Back to Home / #linode / 2009 / 10 / Prev Day | Next Day
#linode IRC Logs for 2009-10-26

---Logopened Mon Oct 26 00:00:46 2009
00:03-!-Berto [] has joined #linode
00:05-!-philoye [] has quit [Quit: philoye]
00:06-!-philoye [] has joined #linode
00:09-!-walterheck [] has joined #linode
00:15-!-philoye [] has quit [Quit: philoye]
00:23<@pparadis>minor library maintenance in a couple of minutes; there will be a brief outage for
00:24-!-Auzy [~kvirc@2002:7a6b:8d21:0:c08e:e8:f3d2:c2e2] has quit [Quit: When two people dream the same dream, it ceases to be an illusion. KVIrc 3.4.2 Shiny]
00:27<Clorith>ono, just this second when I was gonna click it!
00:27<Clorith>damn you pparadis!
00:27*Clorith trots off to buy some .as domains then
00:27*Clorith growls
00:27<Clorith>still not unlocked...slow ass useless people.
00:32-!-ondrej [~ondrej@] has quit [Remote host closed the connection]
00:39<linbot>New news from forums: Webserver Load Roundup Question Session in General Discussion <>
00:43-!-Bohemian [~Bohemian@] has joined #linode
00:43<amitz>I don't understand the , espeically the "Security issues" part of last paragraph. Suppose the root of the transit server is evil, I thought that root still can't automatically able to decrypt the transmission?
00:45-!-walterheck [] has quit [Quit: Ex-Chat]
00:45<amitz>I mean, even if the root of transit server has access to the agent socket, the agent socket of transit server basically just pass around encrypted traffic right? Encryption/Decryption is done at the client and target server right?
00:46<amitz>*has access to the agent socket of transit server,
00:51-!-kassah [] has quit [Ping timeout: 480 seconds]
00:52-!-Berto [] has quit [Remote host closed the connection]
00:55-!-Berto [] has joined #linode
00:59-!-JM [] has quit [Ping timeout: 480 seconds]
01:03<@jed>alright beta testers
01:03<@jed>try your zone files out, and see if it parses them right
01:04<@jed>(you can put in "abc123" for api key or leave it blank, I don't use it yet)
01:05<MJCS>I think I put too much thermal paste on...
01:05<MJCS>261F is not good
01:06<MJCS>There is a problem with this website's security certificate.
01:06<@jed>yeah, there is
01:08-!-panpainter [] has joined #linode
01:09-!-panpainter [] has left #linode []
01:11<@jed>ah, empty zone is a traceback
01:12-!-MJCS [] has quit []
01:12<Jonathan1>empty domain dumps out too
01:12<Jonathan1>valid zone file parses perfectly
01:12-!-Bohemian [~Bohemian@] has quit [Quit: Bohemian]
01:14<Jonathan1>is a cname allowed to point to an IP?
01:14<Eman>why would you point a cname at an ip?
01:14<@jed>did it accept it?
01:15<@jed>(and no)
01:15<Jonathan1>yeah - I got:
01:15<Jonathan1>I only put the ip in
01:15<@jed>I'd think named-checkzone would bitch about that
01:15<Jonathan1>space after the mail
01:15<Jonathan1>mail 123. etc
01:15<@jed>oh, that behavior is actually, believe it or not, correct
01:16<Jonathan1>fairy enough, don't know enough about dns
01:16<@jed>since the other side of a CNAME is supposed to be an absolute name, if you put in it assumes you mean$ORIGIN
01:16<@jed>since$ORIGIN would be legal
01:16<Jonathan1>just wondering if it'd be better to warn of the error than set it up incorrectly
01:16<@jed>well, hard to catch that one
01:16<@jed>someone might genuinely mean
01:17<@jed>I'm trusting named-checkzone to catch everything right now, I'll have to add exceptions (like that one, I'll debate it) as they pop up
01:17-!-litwol|mac [] has joined #linode
01:18<Jonathan1>@ NS - parses fine
01:18<Jonathan1>leave out the final dot, and it traces
01:18<@jed>shouldn't trace, named-checkzone should whine
01:18<@jed>lemme see
01:21<Jonathan1>not sure if this is an issue or not, similar to the ip thing:
01:21<@jed>alright, missing args is fixed
01:21<Jonathan1>mail CNAME - is fine
01:21<Jonathan1>miss the dot, and it adds $origin on
01:21<@jed>yeah, same story
01:21<@jed>anything not fully-qualified gets $origin glommed on
01:21<@jed>that would be an error in your zone before you paste it in my box :^)
01:22<@jed>Jonathan1: just changing @ NS to @ NS traces for you?
01:23<Jonathan1>let me double check, i was doing loads of changes, i'll start fresh with single edit
01:23<@jed>yeah, not for me -- in that case it's just adding $ORIGIN as before, which doesn't matter because I'm ignoring NS records outright
01:23<Jonathan1>nah, I must have done something else sorry, it's fine
01:26<Jonathan1>everything else errors out nicely
01:26<@jed>it's still really rough
01:26<Jonathan1>and would be a faulty file in the first place
01:26<@jed>yeah, exactly :)
01:27<@jed>I'm verifying very little, because if your zone file is broken before you plug it in there, that's not my fault :)
01:27<@jed>named-checkzone is really handy, the Linode Manager uses it too
01:27<Jonathan1>I get your point, but I think if you accept it, it becomes your issue (in general app design)
01:28<@jed>well, it's going to list every API action that it's going to attempt (what I'm writing now)
01:28<@jed>and the API is liable to vomit on something, too
01:28<@jed>in which case I'm going to have it fail cleanly and back out and delete its partial zone
01:29-!-elhippo [] has quit [Quit: Leaving]
01:29<Jonathan1>sounds safe enough
01:29-!-Bohemian [] has joined #linode
01:30-!-Berto [] has quit [Remote host closed the connection]
01:34-!-Berto [] has joined #linode
01:35-!-Bohemian [] has left #linode []
01:36-!-dvyjones_ [] has quit [Quit: Lost terminal]
01:36-!-kelvinq [] has joined #linode
01:37-!-A-KO^ [] has joined #linode
01:38-!-A-KO^ [] has quit [Read error: Connection reset by peer]
01:41-!-A-KO [] has quit [Ping timeout: 480 seconds]
01:42-!-Syrogen [~asdas@] has joined #linode
01:43-!-Trystan [] has quit [Ping timeout: 480 seconds]
01:44-!-BostonWealth [] has joined #linode
01:45-!-Trystan [] has joined #linode
01:48-!-Guspaz [] has quit [Read error: Connection reset by peer]
01:52-!-A-KO [] has joined #linode
02:00<BostonWealth>Yesterday someone asked me to post trace route results to look into latency issue for a client in New Zealand. Here are the results
02:00-!-compwhizii [] has quit [Ping timeout: 480 seconds]
02:05-!-neh [] has joined #linode
02:09-!-blognewb_ [~blognewb@] has joined #linode
02:15-!-blognewb [~blognewb@] has quit [Ping timeout: 480 seconds]
02:25-!-kassah [] has joined #linode
02:36-!-atambo [~atambo@ATAMB0.RES.CMU.EDU] has quit [Remote host closed the connection]
02:42-!-neh [] has quit [Remote host closed the connection]
02:48-!-bogdanbiv [~quassel@] has joined #linode
03:04-!-BostonWealth [] has quit [Quit: CGI:IRC 0.5.6 (2005/02/09)]
03:12-!-arooni-mobile [] has quit [Ping timeout: 480 seconds]
03:15-!-A-KO^ [] has joined #linode
03:17-!-zack [] has quit [Quit: zack]
03:19-!-grawity [] has joined #linode
03:20-!-A-KO [] has quit [Ping timeout: 480 seconds]
03:21-!-A-KO^ is now known as A-KO
03:33-!-ph^ [~ph^@] has joined #linode
03:38-!-kassah [] has quit [Ping timeout: 480 seconds]
03:41<amitz>BostonWealth: the traceroute seems acceptably responsive, you can't make it better. I guess the problems are what had been mentioned in this channel, like unnecessarily large images.
03:42-!-arooni-mobile [] has joined #linode
03:43<amitz>responsive for NZ -> US connection.
03:45-!-ph^ [~ph^@] has quit [Remote host closed the connection]
03:47-!-ph^ [~ph^@] has joined #linode
03:50-!-kelvinq [] has quit [Quit: kelvinq]
04:01<amitz>heh, first time I heard such thing but I dont know if it's true or not. A company doesn't want to pick up the phone in fear of lightning while in a storm..
04:04<chesty>!f z=z
04:04<linbot>chesty: True
04:04<chesty>amitz: ^^
04:05<amitz>chesty: yeah, just googled. Interesting, I will not be on the phone anymore when storm hits.
04:10-!-Syrogen [~asdas@] has quit []
04:10<amitz>heh, the peril of manning a call center :-p
04:14<praetorian>!f a=2 b=3 a+1=b
04:14<linbot>praetorian: Results not found. I'm pretty sure straterra ate them
04:14<praetorian>i'd concur.
04:16-!-bluetegu [] has quit [Ping timeout: 480 seconds]
04:17-!-blognewb_ [~blognewb@] has quit [Quit: Client Quit]
04:20<amitz>say, how to create a dynamic subdomain that dynamically point to a dynamic IP address? Basically I want to contact a PC having a dynamic IP. I want the PC to update the DNS record of the subdomain automatically. Basically I setup a DNS server like bind? Or there is a simpler approach?
04:20<grawity>CNAME it to
04:21<amitz>without relying to dyndns? So maybe a subdomain of a domain I control?
04:22<StevenK>amitz: The simplest way is to run an instance of bind
04:23<amitz>StevenK: I fear that answer but if that's what simple, so be it. Amitz sigh deeply.
04:23<amitz>ooh, or use linode's DNS manager API! Forgot that one :-)
04:24<StevenK>That's simpler again
04:24<StevenK>How often do you expect IPs to change?
04:25<amitz>StevenK: I will probably use the connection once overy week, so the changes that matter is once a week. You have something in your mind?
04:25<praetorian>or use some dnssec
04:25<praetorian>if you control the dns
04:28<StevenK>amitz: Just propagation delays
04:29<amitz>oh. will google about dnssec, will probably have my own DNS since using linodes DNS manager API means that the PC will store my linodes API key..
04:31<chesty>you could, for example, hit a php script on your linode that does the update
04:32<chesty>but linode update their dns every 15 minutes
04:33-!-mcinerney [] has joined #linode
04:33<amitz>chesty: oh, an average of 7.5 minutest + propagation, that's too long.. hmm, let me think
04:35<amitz>maybe I should let the PCs automatically connect to my server.. then I reverse tunnel. Fewer point of failures.. I'm sounding my thought, feel free to feed back.
04:40-!-Strifariz [] has joined #linode
04:41-!-ph^ [~ph^@] has quit [Remote host closed the connection]
04:41-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:29a1:8df4:67f4:778e] has joined #linode
04:48-!-JoeK [] has quit [Ping timeout: 480 seconds]
04:49-!-ph^ [~ph^@] has joined #linode
04:54-!-JoeK [] has joined #linode
05:06-!-JoeK [] has quit [Ping timeout: 480 seconds]
05:09<praetorian>need to get my website to be as good as (design wise)
05:09<praetorian>[they redid the site]
05:11*Strifariz waves baibai geocities
05:11-!-JoeK [] has joined #linode
05:12<Strifariz>I just found out about xkcd's design too
05:12<Strifariz>I really love the broken image links etc
05:12<Strifariz>it's just like old times
05:12<Peng_>Oh my god
05:16-!-Deckert_ [] has quit [Ping timeout: 480 seconds]
05:17<Strifariz>it is BRILLIANT
05:17-!-kelvinq [~kelvinq@] has joined #linode
05:17<Strifariz>me and my friends are looking at our old geocities sites before they go kapoof
05:19<grawity>(...When it works, of course >_> )
05:24-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:29a1:8df4:67f4:778e] has quit [Quit: When two people dream the same dream, it ceases to be an illusion. KVIrc 3.4.2 Shiny]
05:32<Strifariz>so not now? :P
05:33-!-jas4711 [] has joined #linode
05:34<Strifariz>hooray for compiling code excuse for browsing random websites during work hours
05:48-!-chris2 [] has joined #linode
05:49-!-pparadis_ [] has joined #linode
05:49<chris2>Does linode use ecc registered dimms. Also what speed hard drives are used?
05:51-!-pparadis [] has quit [Ping timeout: 480 seconds]
05:55<Strifariz>any of you use ImageMagick?
05:56-!-chemosh [] has joined #linode
05:57-!-chris2 [] has quit [Ping timeout: 480 seconds]
05:58-!-kelvinq [~kelvinq@] has quit [Quit: kelvinq]
06:04-!-Thorgrimr [] has quit [Remote host closed the connection]
06:04-!-HoopyCat [] has quit [Remote host closed the connection]
06:04-!-straterra [] has quit [Remote host closed the connection]
06:05-!-HoopyCat [] has joined #linode
06:05-!-straterra [] has joined #linode
06:05-!-jspiros [] has quit [Remote host closed the connection]
06:05-!-jspiros [] has joined #linode
06:05-!-Thorgrimr [] has joined #linode
06:09-!-fvoice [] has joined #linode
06:10-!-fvoice [] has quit [Remote host closed the connection]
06:17<Peng_>Oh, crhs2 left. No fun.
06:17<Peng_>Err, chris2. This is why tab-complete is good. :P
06:26-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:c977:ec8f:59b0:53fc] has joined #linode
06:27-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:c977:ec8f:59b0:53fc] has quit []
06:29<chemosh>Anyone here with a linode @ atlanta57?
06:46-!-Keith-BlindUser [] has joined #linode
06:46<Keith-BlindUser>!dig txt
06:46<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 14440 IN SOA 2009102569 7200 7200 1209600 14440
06:46<linbot>Keith-BlindUser: [spf] none: does not designate permitted sender hosts (spf record: )
06:48<pparadis_>sorry about the current issues; we're working with the DC on it.
06:49<BarkerJr> has no TXT record
06:50<linbot>New news from forums: Atlanta network issues in System and Network Status <>
06:51<BarkerJr>there you go, chemosh
06:53<Strifariz>not really relevant for here...but anyone know how to do a domain forward in IIS in server 2003? :P
06:54<Keith-BlindUser>What I don't understand, is why my SPF record isn't working! I updated the SAO record and even have a valid email address for it. But it doesn't work!
06:54<HoopyCat>(secretly, it's because atlanta is really done with server-side includes on a geocities page)
06:54<Keith-BlindUser>If we do:
06:54<Keith-BlindUser>!dig txt
06:54<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 13935 IN SOA 2009102569 7200 7200 1209600 14440
06:54<HoopyCat>Strifariz: i'm going to say "yes, but i haven't done it in yeaaaaars, so all i can say is 'it's somewhere in the properties'"
06:55<Keith-BlindUser>We see, that is used. That email does exist. But why it's not working far as an SPF goes I don't know.
06:55<pparadis_>Strifariz: perhaps this is useful?
06:56<Keith-BlindUser>So: I'm confused. Apparently my domain doesn't do SPF or something's just improperly configured, but how that's possible, I don't know..
06:57<BarkerJr>I don't see any TXT record on your domain
06:58-!-cpg [] has quit [Quit: cpg]
06:59<Keith-BlindUser>There is a txt record, but perhaps improperly configured? Value of witch is: v=spf1 ip4: ~all
07:00<HoopyCat>... good morning, this is #linode 1010 AM absecon, #linode 88.1 FM absecon, #linodx 89.7 FM atlantic city, and #linode HD 1 absecon... it's a beautiful day, because today is day 1 of our fall membership campaign, and we're getting things going with a two-for-one challenge from dirk winchester of pound ridge here at 222-1010, that's 222-1010, and at the dollar-a-day level you'll receive a shiny Linode 540 t-shirt...
07:01<pparadis_>. . .
07:01<HoopyCat>ugh, sorry, top of the hour hits and i start regretting the clock radio
07:02<Keith-BlindUser>pparadis_ Does that look incorrect? v=spf1 ip4: ~all
07:02<Keith-BlindUser>That's the contents of the current txt record for
07:02<pparadis_>Keith-BlindUser: you may find this useful:
07:02<Keith-BlindUser>But if we do something like: !spf we get an error. :D
07:02<Nivex>HoopyCat: wow, they really are the same everywhere. One of our stations here just finished up their pledge drive last night.
07:03<chemosh>cheers pparadis_
07:03<Keith-BlindUser>Alright, hold on! :)!
07:03<BarkerJr>keith: you have no txt record at all.. it's not malformed, it just doesn't exist
07:04<BarkerJr>"host -tTXT" returms " has no TXT record"
07:06<HoopyCat>Nivex: i'm always somewhat annoyed at what's being lopped out of morning edition, but, well, i can't complain TOO much
07:07<HoopyCat>Nivex: and i suspect -- although i've never been able to confirm -- that they take different chunks out of each loop of ME, so if you listen long enough, you won't miss anything
07:07-!-bob2 [] has joined #linode
07:07<BarkerJr>I refuse to contribute during these events cause I don't like them
07:08-!-supine [] has joined #linode
07:08<BarkerJr>I don't like how they replace the weekly shows with pledge week versions
07:08<BarkerJr>but that's ok... I can pull those up on their podcasts
07:09<BarkerJr>then donate directly to your favourite shows if they're independent
07:15<Keith-BlindUser>!dig txt
07:15<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 12707 IN SOA 2009102569 7200 7200 1209600 14440
07:15<Keith-BlindUser>Now we need to wait for the zone generation. ;)
07:17-!-chemosh [] has quit [Remote host closed the connection]
07:18-!-kelvinq [] has joined #linode
07:19<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: 14440 IN A | ;; AUTHORITY SECTION 14440 IN NS 14440 IN NS 14440 IN NS 14440 IN NS
07:19<HoopyCat>phew, the car started
07:19-!-Deckert [] has joined #linode
07:20<Keith-BlindUser>Let's see..
07:23<HoopyCat>kinda missed the whole "closing all the doors" thing when we went to the coffee shop this weekend, so the battery was a little flat... was out there last night to prod the computer, and the voltage was so low the LCD was unstable. was quite relieved this morn when the DC converter kicked in and the taillights went to full brightness.
07:25<Keith-BlindUser>!dig txt
07:25<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 12073 IN SOA 2009102569 7200 7200 1209600 14440
07:30-!-Deckert [] has quit [Read error: Connection reset by peer]
07:31-!-Deckert [] has joined #linode
07:37-!-sc0field [~rajiv@] has joined #linode
07:37-!-sc0field [~rajiv@] has quit []
07:37<BarkerJr>don't you hate when people on the radio say "insuring" rather than "ensuring"?
07:37<Keith-BlindUser>!dig txt
07:37<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 11344 IN SOA 2009102569 7200 7200 1209600 14440
07:38<Keith-BlindUser>TXT records exist, so why is Linbot not returning anything?
07:38<bob2>doesn't exist
07:39<bob2>at least on
07:43<Solver>check the SOA
07:46-!-sc0field [~rajiv@] has joined #linode
07:47<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: 12720 IN A | ;; AUTHORITY SECTION 12720 IN NS 12720 IN NS 12720 IN NS 12720 IN NS
07:48-!-bryanl [] has quit [Quit: Leaving...]
07:49<BarkerJr>why does it say NOERROR?
07:50<praetorian>there was no error!
07:51-!-quanin [~james@] has joined #linode
07:56<BarkerJr>so, does not exist isn't an error
07:56<Solver>indeed :)
07:57<Keith-BlindUser> interesting. Maybe the master DNS hasn't yet updated; but txt records do exist.
07:57<Solver>I just checked and each NS has the same SOA
07:57<bob2>perhaps linode's dns stuff is broken
07:58<bob2>or you put the record on another name
07:58<Solver>but perhaps there is a lag before new entries are pushed to the DNS by linode
07:58<Solver>I run my own name servers so no idea
07:58<Keith-BlindUser>How many IPS are you using? I suppose you can't just run an DNS server from one.
07:59<Solver>you should have at least 2 nameservers
07:59<bob2>non sequitor
07:59<Solver>I have 2 linodes and they are indeed my 2 nameservers
07:59<bob2>having two ips on the one vps doesn't count as two nameservers
07:59<praetorian>bob2: schadenfreude
07:59<Solver>nameservers should be physically and logically seperated
08:00<praetorian>i only have 8 ns's
08:00<BarkerJr>seperate networks
08:00<Solver>RFCs recommend a max of 7 :) really
08:00<bob2>anyway, since linode is happy to slave to your nameserver, this is easily solved
08:00<praetorian>Solver: well when one is down. it will be perfect! ;)
08:00<Solver>praetorian: hahaha :)
08:01-!-ubuntuisloved [] has quit [Ping timeout: 480 seconds]
08:01<Solver>bob2: very true
08:01<linbot>Keith-BlindUser: [spf] none: does not designate permitted sender hosts (spf record: )
08:01*Solver is running his own as he wants to play with dnssec
08:01*Keith-BlindUser groans.
08:01-!-jonny5 [] has joined #linode
08:02<jonny5>woo linode resize
08:03<praetorian>re :
08:03<praetorian>did anyone "view source" ?
08:03<Trystan>not me
08:04<grawity></TABLE FONT="800px"> is gold.
08:04<praetorian>? :-)
08:06-!-Keith-BlindUser [] has quit [Quit: Freedom Chat - Your Home Away From Home | | tIRC script by the Freedom Chat Leets]
08:07<jonny5>Anyone know how long disk resizing usually takes?
08:07<@irgeek>Depends on a lot of factors.
08:07<@irgeek>Resizing up is usually faster than resizing down.
08:07<Solver>upsize me!
08:07<jonny5>Just thought it'd be instant so was a bit scared it was taking a while :)
08:08*Solver recommends a backup before any resize, even if it 'should' work :) I'm a sysadmin, it's my job to be distrustful of technology ;)
08:09<jonny5>Good advice
08:09<jonny5>woop woop it is finished
08:09<@irgeek>There's no way to get progress information out of the resize process so it just sits there until it's done.
08:11<jonny5>I hear Slicehost do instant resizes but they also sound like smell
08:15-!-jonny5 [] has quit [Quit: byebye]
08:18-!-arooni-mobile [] has quit [Ping timeout: 480 seconds]
08:20-!-SubZero [] has joined #linode
08:22-!-Deckert [] has quit [Read error: Connection timed out]
08:25-!-awnstudio [] has quit [Quit: HydraIRC -> <- Would you like to know more?]
08:33-!-bogdanbiv [~quassel@] has quit [Ping timeout: 480 seconds]
08:41-!-BostonWealth [] has joined #linode
08:43<BostonWealth>I was told the other day to post traceroute result tests from a client in New Zealand trying to access my site to investigage latency issues; here are the results
08:44-!-JM [] has joined #linode
08:45<JshWright>208 ms from NZ to Texas? Doesn't seem that bad to me
08:47-!-linville [] has joined #linode
08:48<BostonWealth>JshWright.. the client says it takes a long time
08:49<quanin>don't think it's a latancy issue though. not if 208 MS is the longest it takes for a traceroute packet to get from one end to the other.
08:50<Trystan>JshWright: it isnt to bad
08:50<Trystan>i get around the same from Australia
08:50<G>from exp, 200ms anywhere from NZ is good :)
08:51<G>sometimes you can only get 200ms down the road ;)
08:51<Trystan>their networks arnt that bad
08:51<Trystan>(though, i note the client is on telstra clear... thats... bad)
08:51<G>Trystan: TelstraClear is okay in NZ
08:52<Trystan>telstra clear is ONLY in NX
08:52<Trystan>so my comment was related to them there :)
08:53<G>Trystan: I was assuming your comment was because of the 'Telstra' there
08:53<Trystan>partially :)
08:53<Trystan>Telstra Clear are meant to be better organised than
08:54<Trystan>since, they are actually in the oposit position
08:54<BostonWealth>Hi quanin. Good to see you. What could the issue be then. The site images, expire headers, script, etc. have been optimized
08:54<G>the 34ms to TelstraClear is pretty good, and the delay to reach is okay
08:54<G>BostonWealth: client's speed?
08:55<BostonWealth>HI G, should I ask him that question?
08:55<quanin>BostonWealth: i would. especially if everything else checks out okay.
08:55<BostonWealth>Will do and report back!
08:55<G>quanin: askk him if he's on the Cable or ADSL too
08:55<Trystan>BostonWealth: any other Kiwis having the same problem?
08:56<BostonWealth>Good question Trystan.. no he is the only one
08:56<spkitty>it's his connection
08:56<Trystan>latency, especially in island countries like NZ and Aus
08:56<spkitty>also the site hasn't been optimised
08:56<Trystan>are reliant on cables between here and the US
08:56<quanin>BostonWealth: then yes, it's his connection.
08:56<spkitty>still a lot to gain in scripts and images
08:56<spkitty>you're using .jpgs for godsake
08:57<G>Trystan: you do realise that NZ is actually in the middle of one of the highest capacity pacific cables right....
08:57<BostonWealth>my web min says he can't use your programs
08:57<Trystan>so we tend take the same paths
08:57<Trystan>G: still going over the same pipe.
08:57<quanin>BostonWealth: ... then you need a new web admin. they're not hard.
08:57<G>Trystan: right, but it's AU<->NZ<->US (although some AU traffic can skip NZ)
08:57<Trystan>G, exactly
08:58<G>thats the Southern Cross Cable
08:58<spkitty>yeah your web admin is useless
08:58<Trystan>AU and NZ traffic tends to go over the same cable
08:58<Trystan>if its working fine for the majority of AU and NZ customers, chances are is local to the specific client
08:58<G>BostonWealth: but anyway, if he's on the TC Cable Network, I'd be pretty surprised if he's really having issues
08:59<quanin>ZOMG it's not instant and it's all your fault! die!
08:59<G>BostonWealth: the ADSL network is still for TC pretty much poor, because it goes into Telecoms exchanges
08:59<Trystan>and, alot of our traffic can skip NZ now
09:00<BostonWealth>here is the explanation why he is not using
09:00<BostonWealth>My reasoning behind not combining all the CSS sprites is because I did not design your template and it would take a long time to rewrite it to properly use CSS sprite (like 4+ hours) and since images are now being caches, this will only save a little bit time the very first time a user accesses your site. After that they will access the cached images and there won't be a delay
09:00<BostonWealth>Is the fact that he is using Windows XP make any difference
09:00<G>BostonWealth: remember that it relies on their Web Browser to cache properly
09:01<spkitty>he's wrong by the way
09:01<spkitty>cache's are not served up on a per user basis
09:01<spkitty>once somebody has accessed a page it is cached for all
09:01<quanin>roughly translated, it's too much like work and your web admin's alergic.
09:01<spkitty>you can bring the cache size down by not having stupid sized images though
09:01<Trystan>quanin +1
09:01<spkitty>so basically, he's an idiot
09:02<BostonWealth>quanin.. actually he is doing a great job
09:02<Trystan>arguable given the above reasoning
09:02<quanin>BostonWealth: no offense, but... if he were my web admin, i'd of shot him by now.
09:02<spkitty>very arguable
09:02<BostonWealth>quanin.. take a look at my site and tell me what YOU would do. I value your opinion
09:03<spkitty>my offer still stands if you're willing to pay me part upfront ...
09:03<quanin>without even looking at your site, i can tell you right now.
09:03<quanin>ditch the .jpg's, as was already said.
09:03<BostonWealth>replace with what quanin
09:03<spkitty>i told you this 5 times
09:03<quanin>.gif, or if you absolutely as in no other choice have to, .png. just... not .jpg.
09:03<BostonWealth>is there a way to limit what contributors to wordpress use.. I have almost 20 contributors who post
09:04<BostonWealth>quanin do you recommend a program to convert jpg to gig
09:04<spkitty>you don't do that
09:05<BostonWealth>Quanin this is what he said: John were to use my Gimp technique he would save a lot of space without any image degradation
09:05<BostonWealth>so quanin do you recommend gimp?
09:06<BostonWealth>so how to change out of exiting jpg then
09:06<spkitty>are these the original images?
09:06<spkitty>where do they come from, what's the *original* format and size
09:06<BostonWealth>the orginal images are all jpeg yes
09:06<Trystan>dont have the ORIGINAL images?
09:06<spkitty>no, are *those* the original images
09:06<Trystan>in like.. .png?
09:06<BostonWealth>I will ask
09:06<Trystan>that isnt the format i meant.
09:07<spkitty>.psd ?
09:07<Trystan>i had a mental blank and couldnt think of it either lol
09:07<spkitty>if he's using gimp then i doubt it
09:07<spkitty>gimp is a poor replacement for a real photo editing tool
09:07<Trystan>yea, but that guy didnt do the original design from what BostonWealth said earlier I geather
09:08-!-grawity [] has quit [Ping timeout: 480 seconds]
09:08-!-row [] has quit []
09:08<BostonWealth>why is gimp bad and what would you recomend for a real phote editing tool
09:08<quanin>judging by what i helped this guy with a while ago, sounds like the web admin kind of followed sections of like 6 different user manuals. and got about 5 of them wrong.
09:09<BostonWealth>Yes the web admin is relativley new, two month
09:09<spkitty>not gimp or anything else
09:09<BostonWealth>quanin yes, thanks for all the help you provided in the past; you are a valuable member here
09:10<BostonWealth>any particular photoshop.. so make the images into gif
09:10<quanin>BostonWealth: not really. i've just broken enough things in the past to actually know what *not* to do. and he's pretty much done it all.
09:10<Trystan>if you dont have originals
09:10<Trystan>you probably want the images remade
09:11<spkitty>what ... any particular photoshop ... oh god
09:11<BostonWealth>Trystan, what is the best method to remake the images.. thank u
09:11<spkitty>Trystan: he hasn't explained how he gets the images yet, he has new ones for each blog post obviously so he'd have to have originals somehow
09:12<spkitty>BostonWealth: is english your first language?
09:12<BostonWealth>quanin how do you put someone on ignore please
09:13<spkitty>... whatever
09:13<praetorian>BostonWealth: /ignore nick
09:13<spkitty>he wont understand praetorian
09:13<BostonWealth>thanks praetorian
09:13<spkitty>he couldn't even figure out how to change his nick last time
09:13<quanin>rm -rf /broken/system
09:13<BostonWealth>wow spkitty ignored great
09:14<BostonWealth>do I have to ignore for every new session
09:14<quanin>it depends on your client.
09:14-!-bogdanbiv [~quassel@] has joined #linode
09:14<BostonWealth>quanin what was that rm-rf/broken/system for please
09:14<spkitty>only if i keep changing so that i can tell you what a broken piece of shit you are
09:14<Trystan>BostonWealth: to be frankly blunt. Sounds like you need to employ someone to do Graphic Design, and either chat with your web admin to tell them exactly what you want and if it isnt achieved, perhaps move on
09:16<spkitty>Trystan: not sure how he'll find anybody when he insists they do the work upfront before they get paid
09:16-!-sbruner [] has joined #linode
09:16<BostonWealth>Trystan: one more question on the subject please.. he already told me it may be too much for him, what qualities do I ask someone regarding Graphic Design.. what do they need to know how to do, and can you setup wordpress to optimize and accept only certain types of images
09:16-!-ubuntuisloved [] has joined #linode
09:16<pparadis_>spkitty: please keep it civil.
09:16<spkitty>sorry pparadis_ being trying to get it into this guys head for 2 days now what he should be doing, and he has 0 clue
09:17<pparadis_>we all start somewhere, and some pairings of minds work out better than others.
09:17<Trystan>I have no idea with wordpress, not familiar with it. Re: a Graphic designer, ask to see their portfolio, and discuss with them what tools they use and what they would give you.
09:17<BostonWealth>pparadis.. spkitty is one of the reasons along with blind keith that make this chat really bad and leaves Linode with a bad experience .. read the chat from a few days ago.. it is bad for Linode business model
09:17*spkitty rolls eyes
09:17<sbruner>pressed shutdown on linode console. how long until it shutsdown?
09:17<BostonWealth>thanks Trystsn
09:18<spkitty>sbruner: it will be in the jobs task on your dashboard
09:18<pparadis_>spkitty / BostonWealth: i understand that there's tension here, but let's try to keep it out of the channel.
09:18<sbruner>it just shutdown
09:18<BostonWealth>pparadis take a look at this please and tell me what you think regarding a client in NZ with latency issues
09:18<sbruner>to restart just BOOT debian?
09:19<BostonWealth>pparadis.. I am always very polite and civil, the name calling has not been from my side or the insults hurling
09:19<spkitty>just click the boot button sbruner
09:19<pparadis_>i'm just trying to remain neutral here as a company representative.
09:19<sbruner>spkitty: thanks
09:20<BostonWealth>pparadis. someone suggested changing from dallas to sacramento
09:20<BostonWealth>but most of my clients are state side
09:20<G>BostonWealth: as we said, it's pretty a-typical of NZ->US
09:20<G>BostonWealth: it's likely to be speed issues, not latency
09:22<BostonWealth>I have asked John to join us to discuss the type of images he starts with
09:24<praetorian>i most concur. those traceroutes are what i'd expect.
09:24-!-mortie [] has joined #linode
09:25<Trystan>you look familiar praetorian..
09:25-!-orudie [] has joined #linode
09:26<praetorian>most? must.
09:26<Trystan>but I dont know which one of the two I know you are.
09:26<Trystan>One from another IRC network
09:26<Trystan>another used to work for an Aus ISP
09:27<praetorian>oh, not the second one ;)
09:27<Trystan>answers that :)
09:27<Trystan>Yea, I remember that was the first thing I asked you over there lol
09:28<BostonWealth>Hi Mortie.. how do you start off with your images, what format are they in
09:29-!-mortie [] has quit [Remote host closed the connection]
09:29<Peng_>...Bye mortie.
09:29-!-mortie [] has joined #linode
09:30<Peng_>mortie: wb
09:31<BostonWealth>yes we see you mortie
09:32-!-Desph [] has joined #linode
09:32<BostonWealth>I was asking earlier how do you start off with your images, what format are they in and what tools do you use for the images
09:32<BostonWealth>peng.. hi! what is wb
09:33<Trystan>welcome back
09:33<BostonWealth>quanin there is your answer
09:33<BostonWealth>mortie, the folks here are saying we should not be using jpg
09:33<BostonWealth>slows down the server
09:34<spkitty>mortie is the best invisible friend
09:34<BostonWealth>the experts here say jpg slows the server
09:34<straterra>Experts here?
09:34<Trystan>its a lie.
09:34<chesty>i'm here
09:34<BostonWealth>yes like quanin, amitz, etc.
09:35<JshWright>jpgs won't slow ther server, they're just larger than they need to be for most applications
09:35-!-sbruner [] has quit [Remote host closed the connection]
09:35<Trystan>it wasnt actually said that they slow the server
09:35<JshWright>small graphics should be pngs or gifs
09:35<Trystan>that was
09:36-!-Stas [] has joined #linode
09:36<BostonWealth>does anyone know if wordpress automatically converts to jpg when images are added
09:36<JshWright> <- First google hit for "png jpg gif", I haven't read it but at a quick glance it doesn't look like they're smoking crack
09:37<straterra>BostonWealth: I..don't know why it would
09:37<spkitty>it doesn't straterra
09:37<JshWright>BostonWealth: there's probabably a plugin that does that, but it would seem pretty pointless
09:37<BostonWealth>mortie.. I just checked the images in the image library and all your images are put in png
09:37<JshWright>BostonWealth: I'd recommend doing some reading on basic web design
09:38<JshWright>You might also want to check out some of these resources: (some of them are way more advanced than you need though)
09:38<fapestniegd>my fremont linodes can't ping any of my NJ linodes. Traceroutes drop 6-hops out at
09:38<fapestniegd>(if you didn't already know)
09:39<BostonWealth>Mortie can you convert them to png and send them to me and I will replace them or we can email about it later
09:39<linbot>JshWright: [mtr] 10 hops, no loss, last hop average RTT was 81.3ms (These results brought to you by urmom and the number 42)
09:39<spkitty>somebody tell BostonWealth that converting them to png isn't going to make them any better quality
09:39<Trystan>converting them to png isn't going to make them any better quality
09:39<BostonWealth>jshwright.. thanks much for the links.. I have them bookmarked for reserch
09:40<straterra>BostonWealth: converting them to png isn't going to make them any better quality
09:40<JshWright>BostonWealth: you also may want to consider paying someone who knows that they're doing to do this for you
09:40<BostonWealth>no.. I was told here that jpg was the problem with our speed issue
09:40*straterra head-desks
09:40<JshWright>BostonWealth: what's the site in question?
09:40<spkitty>know you guys know how i feel
09:41-!-row [] has joined #linode
09:41*straterra is afraid to click
09:41<BostonWealth>mortie.. yes thanks I will look at them
09:41<spkitty>it is only a matter of time before you too are /ignore'd for offering him advice he doesn't agree with
09:41<linbot>fapestniegd: [mtr] 7 hops, ???: 100.0%/0.0ms (These results brought to you by urmom and the number 42)
09:41<Peng_>spkitty: I think it was more the "being a jerk" than the advice.
09:41*Peng_ gets some gasoline.
09:41<fapestniegd>correction, atlanta.
09:41<straterra>Page loaded quickly for me..
09:41<spkitty>actually i was serious when i asked if english was hi first language Peng_
09:42<spkitty>it's the way he writes and doesn't seem to understand what we say, thought maybe if he spoke another language better somebody could help
09:42-!-johndbritton [] has joined #linode
09:42*fapestniegd needs more coffee
09:42<Trystan>straterra: it looks to be a number of issues assimilating to cause the particular client their problem
09:42<Peng_>spkitty: Oh, that was what got you /ignored? I agree; I'd just never had had the nerve to ask. :P
09:42<JshWright>BostonWealth: you should also install Firebug and Yslow
09:42<adj>important: parallellize downloads across hostnames; useful: leverage browser caching, minimize DNS lookups, optimize images, remove unused CSS, serve static content from a cookie-less domain, minimize cookie size, use efficient CSS selectors
09:43<BostonWealth>jshwright. firebug and yslow installed grade 86 B
09:43<BostonWealth>adj.. yes but 86 B is pretty good no?
09:43<adj>BostonWealth: does the site load within your expectations?
09:43<BostonWealth>for me yes..
09:44<BostonWealth>but for one client complaint.. I am not happy
09:44<straterra>then blame the client..and call it a day
09:44<adj>BostonWealth: a number means nothing. if its good and your conversion rates are good. then to hell with some plugins score
09:44<BostonWealth>we are big on customer service
09:44<Trystan>loads quick for me to from the other side of the world
09:44<Trystan>BostonWealth: you can be as big on customer service as you like if someone else has a problem unrelated to your service theres not much you can do
09:44<BostonWealth>yes but customer service has to consider every client!
09:44<spkitty>but that client is worth 7 million dollars to him!
09:44<straterra>You can't fix their network issues
09:44<spkitty>(yeah right ...)
09:45<BostonWealth>Mortie. thank you .. see you on the other side.. you see my continuing quest for perfection...just hit /quit
09:45<Trystan>you know..
09:45<BostonWealth>mortie just do /quit and you are logged out.. thanks!
09:45<Peng_>Does mortie talk?
09:45<Trystan>only YOU could see mortie talking
09:45<spkitty>i really think mortie is imaginary
09:46<straterra>Unless you want to install fiber for the client..running from their office to yours..there isn't much you can do about it
09:46<Trystan>i assume because he was sending you a personal message
09:46<Trystan>rather than sending it to the channel
09:46<BostonWealth>[mortie(] Ben ~ do you need me anymore? I only speak English ;-)
09:46<Trystan>so everyone is assuming he is imaginary :D
09:46<BostonWealth>did he log in incorrectly
09:46<BostonWealth>then why was he not seen
09:46<Trystan>he just told it to send directly to you rather than everyone else
09:46<Trystan>i can send a message that goes to one specific user
09:46<Trystan>or i can send to the channel
09:46<BostonWealth>so how did he mess up
09:46<Trystan>he was doing the first
09:47<Trystan>no idea.
09:47<BostonWealth>told him to log into irq channel
09:48<BostonWealth>if anyone here is a good graphics designer, feel free to email me at
09:48-!-azaghal [] has quit [Ping timeout: 480 seconds]
09:48<BostonWealth>I pay bills on time, and I can provide a reference of a web admin I found here
09:48<spkitty>he also doesn't pay any money in advance
09:49<spkitty>he wants *you* to do the work for him and *then* he will pay you
09:49<linbot>New news from linodelibrary: Using SSHFS on Linux and MacOS X <>
09:50<BostonWealth>Site says it won't work well with my browser. Let me see if I can change.
09:50<BostonWealth>that is what mortie said when he tried to log in irq chat
09:51-!-SubZero [] has quit []
09:51*G shakes head and just goes to bed
09:52-!-superkuh [] has quit [Remote host closed the connection]
09:53<@jed>break it:
09:53<@jed>if you make it trace I want to know :^)
09:55<pparadis_>bd_ ^
09:55*DephNet[Paul] wonders what this "irq chat" BostonWealth is refering too
09:55-!-Stas [] has quit [Quit: Stas]
09:55<Peng_>DephNet[Paul]: Something with ICQ and IRC? :D
09:55<Peng_>Or interrupts?
09:56<DephNet[Paul]>peng_, could be
09:56<Peng_>jed: Will that be part of the manager?
09:56<adj>irc interuprts me all the time, so that makes sense
09:57<@jed>Peng_: since it's written in Python and uses the API, it'd have to be reworked :)
09:57<@jed>maybe, I won't say yay or nay
09:58<BostonWealth>meant to say irc
09:58-!-ajmitch [] has quit [Ping timeout: 480 seconds]
10:01<Peng_>A typo? That's boring. :P
10:01<Peng_>jed: Well, I guess it doesn't really matter; I just hope people will be able to find it.
10:04<BostonWealth>thanks everyone for the assitance today
10:04-!-BostonWealth [] has quit [Quit: CGI:IRC 0.5.6 (2005/02/09)]
10:05-!-jas4711 [] has quit [Read error: Connection reset by peer]
10:05-!-djg320 [] has quit [Remote host closed the connection]
10:05-!-jas4711 [] has joined #linode
10:06-!-robert [] has joined #linode
10:06<robert>is anyone else aware of the mzima/cavtel routing issue?
10:07<array>robert: atlanta is currently experiencing routing issues:
10:07-!-jas4711 [] has quit [Read error: Connection reset by peer]
10:08<robert>i'll have to learn to check the forums
10:08<robert>thank you array
10:08-!-jas4711 [] has joined #linode
10:09<Trystan>I see how it is.
10:09<robert>oh wait, atlanta hosts my email server :)
10:09<robert>i'll have to learn to check the forums after this is resolved
10:10<straterra>Woah woah woah
10:10<straterra>Whats up with xkcd today?
10:10<Bdragon>geocities shutdown
10:10<Peng_>straterra: What? It's always been like that.
10:10<spkitty>let me guess it is a comic in which megans name is mentioned
10:10<straterra>Peng_: lies
10:10-!-jonny5 [] has joined #linode
10:10<jonny5>Hey all
10:10-!-jas4711 [] has quit [Read error: Connection reset by peer]
10:11<cdlu>no, xkcd itself is the comic today.
10:11<cdlu>it's most excellent.
10:11<Trystan>it is :)
10:11-!-jas4711 [] has joined #linode
10:11<jonny5>Geocities :)
10:12<Bdragon>Reminds me of that one stylesheet for css zen garden ;)
10:12-!-ajmitch [] has joined #linode
10:12<Bdragon>(bruce lawson's)
10:13<jonny5>Hmm my Linode seems really sluggish since resizing to a bigger one. Wonder how to check
10:14-!-robert [] has quit [Remote host closed the connection]
10:17*Pryon wonders, as requested
10:18<jonny5>Thanks for wondering
10:20<Pryon>did you migrate within the same data center? If so, is it atlanta? If not, what else has changed?
10:22<Peng_>jonny5: Sluggish how?
10:22<jonny5>Within the same data center (Freemont), from 720 to 1080. Loading dynamic web pages seems much slower. I'm doing a full check on MySQL databases (can't hurt) and letting it settle down for an hour and see if my Munin graphs show anything after that. I have noticed there is a lot more 'iowait' than before but have to wait until this MySQL check is done to check that properly. I was suspecting maybe this host has someone chewingup the hard drive? brb
10:24<straterra>You could be on a more active host
10:27-!-Guspaz|m [] has joined #linode
10:28-!-MJCS [] has joined #linode
10:31-!-Deckert [] has joined #linode
10:33-!-laser` [] has joined #linode
10:35-!-JongMan [] has joined #linode
10:44-!-Berto [] has quit [Remote host closed the connection]
10:45<Guspaz|m>It looks like Caker is going to pass the first month in a year with no news posts!
10:47-!-azaghal [] has joined #linode
10:48-!-Deckert [] has quit [Ping timeout: 480 seconds]
10:48<MJCS>is he feeling ok?
10:48<Guspaz|m>I hope so. It's been nearly a month and a half since the last news post.
10:49<DephNet[Paul]>perhaps he's saving all the news for a mega post on Halloween
10:49<@jed>actually, what's coming next is fairly awesome
10:49*jed rubs hands together
10:49<jforman>jed: each linode comes with a shark with a laser beam attached?
10:49<Guspaz|m>But is it coming in October? We can't have a news post gap!
10:49<Guspaz|m>Ooh, sharks with frikkin' laser beams attached to their frikkin' heads!
10:50<@jed>no, we sold out to rackspace and got $2m each
10:50<@jed>and linode is closing
10:50<@jed>ha! gotcha.
10:50<DephNet[Paul]>only $2million?
10:50<DephNet[Paul]>i could have got you $10million ;)
10:50-!-TimothyA [~MeepMeep@] has quit []
10:50<@jed>for all 10 of us?
10:50*jed interested
10:50<@jed>do go on
10:51<DephNet[Paul]>well, no
10:51<DephNet[Paul]>i only said YOU ;)
10:51<DephNet[Paul]>the other 9 would share the rest :P
10:52<Peng_>jed: Is it.....IPv6? :D
10:53<@jed>it could be, or it could be the sharks with laser beams thing, or it could be prices cut in half, or it could be a new datacenter on the moon, or it could be bumping everything 1%
10:53<@jed>tough to say, it could be a lot of things
10:53<Peng_>jed: Maybe it's all of them, and more!
10:53<jforman>jed: on the moon? you guys heard we're trying to shoot rockets at the moon....right?
10:53<@jed>plus a pony!
10:53<Peng_>Can I have the laser beams put in my pony instead of the shark?
10:53<straterra>I'm banned from contact with poneys
10:54<@jed>jforman: we have a tactical defense system consisting of bruce willis and a team of oil drillers
10:54<DephNet[Paul]>prices cut in half would be awesome, but a European POP would be awesomer
10:54<jforman>jed: i was hoping for samuel l jackson, you know, take some mf'ing snakes on a plane up there!
10:54<Guspaz|m>I hope it's bumping stuff up, I could care less about IPv6, it's pointless atm since no endusers support it :P
10:55<@jed>(or it could be something I didn't mention which will surprise you all)
10:55<Guspaz|m>Or, sharks with laser beams. That'd be awesome too.
10:55<Peng_>Oh, maybe something with backups. Or KVM!
10:55<spkitty>is that something *all* of those things?
10:55<Guspaz|m>jed: You're starting a lifecam feed?
10:55<Peng_>jed: You bought Rackspace? :D
10:55<Guspaz|m>That'd be pretty surprising.
10:55<jforman>jed: can you guys add a "countdown to the secret awesome" to we are going to have trolls coming in here asking otherwise
10:55<Peng_>Actually, IPv6 would be a problem, cuz I don't know how to combine native v6 with my HE tunnel.
10:55<straterra>what secret awesome?
10:55<@jed>I love messing with you guys's heads
10:55<@jed>it makes this job sooooo worth it
10:55<straterra>Peng_: why combine?
10:56<Peng_>straterra: Native is better, but I want to keep my old IPs up for at least a while.
10:56<jforman>jed: yeah, we're happy to support your drug habit with our reoccuring linode payments... ;)
10:56<straterra>That's easy to do
10:56<Peng_>At least, like, until I can click the button 50 times to change my DNS records. :P
10:56<straterra>Keep the tunnel up..with the IP's..
10:56<@jed>actually here is something that's kind of cool, but not the magic awesome:
10:56<straterra>Then just configure IPv6 addys/routes on the eth0 interface..
10:56<@jed>feel free to break it
10:56<straterra>and you have both
10:56<amitz>heh, I miss the action again. Curse my bad luck these days.
10:56<adj>no need to buy rackspace. just make sure the planet doesnt blow up another transfer
10:56<jess^>morning boys
10:57<JshWright>jed: but that connection is untrusted!
10:57-!-bluetegu [] has joined #linode
10:57<JshWright>(so sayeth firefox)
10:57<@jed>JshWright: heavens above!
10:57<adj>i dont need any more fanatical support (or emails trying to sell me some every couple of hours)
10:57<Peng_>jess^: Looky, DNS manager imports from zone files. ^
10:57<Guspaz|m>For shame, jed uses an invalid cert :P
10:57-!-elfgoh [] has joined #linode
10:57<Peng_>straterra: Well sure, but I'd want to make sure the tunnel didn't get used for outbound traffic by default, and I dunno how to do that. :D
10:57*Peng_ shrugs.
10:57<@jed>it's called "free SSL in 30 seconds with enough effort exerted to push a rock an inch"
10:57<Peng_>And, really, the DNS records would be a nightmare.
10:57<straterra>Peng_: simple..change the default route
10:57*JshWright add an exception, but only because it's jed
10:57<erikh>jed: bullshit.
10:57<erikh>it doesn't take that much effort.
10:57<Peng_>straterra: Oh, that would do it? Nice.
10:58<@jed>erikh: size of the rock is a consideration
10:58<straterra>Eclipse POWAH
10:58<erikh>jed: I still maintain my stance.
10:58<jess^>i have SO MANY things to do this week.
10:58<Guspaz|m>Speaking of Undertow, that was a fun game.
10:58<Guspaz|m>Especially since Microsoft gave it away free.
10:58<JshWright>"Powered by Python 2.6 and nginx." fcgi?
10:58<Guspaz|m>Then I paid for the expansion.
10:58<@jed>JshWright: mmhmm
10:59<jess^>i want to finish my online portfolio etc but it's hard to do when you have self-images issues :/
10:59<erikh>jess^: just streak at the mall
10:59<erikh>cures that right up quick
10:59<erikh>or it puts you in prison
10:59<erikh>one of those two.
11:00<Guspaz|m>That's assuming that prison doesn't also cure it.
11:00<erikh>the issues disappear because the self image does too.
11:01<erikh>see also: the military
11:01*erikh ducks
11:01<jess^>the military won't take me
11:01<Peng_>I wonder if I could change the default outgoing IP address?
11:01<jess^>forcibly or otherwise.
11:02<jess^>i have cerebral palsy, so i'm medically unfit for military duty.
11:02<jess^>you guys can all go off to war. i'll knit socks.
11:02<straterra>oh, cool
11:02<straterra>My state offers a discount for renewing plates online
11:03<Peng_>Ehh, I probably have a long enough list of minor medical conditions to get me out of the military.
11:03*erikh too
11:03<straterra>jess^: are the socks...normal?
11:03<erikh>straterra: they only have room for 3 toes.
11:03<Guspaz|m>jess^: The military has lots of things like desk jobs :P
11:03<jess^>straterra: why wouldn't they be normal?
11:03<straterra>Cause you have cerebral palsy :P
11:04<erikh>straterra: ouch. I commend your complete tastelessness.
11:04<straterra>He's states before problems with driving..and now unable to join the military..
11:04<straterra>It was a serious question on whether he has the precision and dexterity to knit properly
11:04<jess^>Guspaz|m: i have that base covered too. for the most part, i'm a lesbian, i.e. i'm only incidentally heterosexual, i.e. the military doesn't want any of the tounge-in-groove girls in their outfit
11:05<Guspaz|m>I hear that the US military is changing that.
11:05<Guspaz|m>In Canada, you can't enlist in the military, because we have a professional military (as in, it's a profession, you need to apply for a job as a soldier, and many people are turned down)
11:05<adj>canada has a military?
11:05-!-kelvinq [] has quit [Quit: kelvinq]
11:05*straterra shrugs
11:05<jess^>lol, he/she/whatever
11:05<Guspaz|m>I think that you need to pass pretty strict fitness stuff to get in the military.
11:05<jess^>getting your bases covered? :P
11:05*cdlu was turned down by the Canadian military
11:06<Guspaz|m>Failed the fitness test?
11:06<cdlu>didn't get that far but I would have :)
11:06*cdlu applied to RMC in early 1999
11:06<straterra>It's the internet. men are men, women are men and girls are men pretending to be cops
11:06<cdlu>just as well I didn't get in, I'd probably be in Afghanistan now
11:06<Guspaz|m>Isn't the fitness test the first part?
11:07<cdlu>there were a bunch of written tests
11:07<cdlu>and then an interview
11:07<cdlu>the physical was in there somewhere too
11:07<cdlu>the fitness test would have come after the interview, but I reached the end of the road at the interview
11:07<Guspaz|m>I'd fail the physical, no question. I'm not heavy, but I'm not exactly in shape.
11:07-!-neh [] has joined #linode
11:07*cdlu weighs an eighth of a ton
11:08<erikh>cdlu: ouch.
11:08<cdlu>which is a fancy way of saying 250 pounds :)
11:08<cdlu>but I wasn't quite that big when I applied
11:08<spkitty>i don't know how much i weigh in your moon units
11:08<Guspaz|m>275 if you're exactly an eigth :P
11:08<cdlu>Guspaz|m, you must have strange tons. :)
11:08<Guspaz|m>Although, if you weren't talking metric tons, then I'll just throw up my hands at the crazy Imperial system.
11:09<cdlu>I convert tons to pounds tonnes to kilos. Take your pick. :)
11:09<Guspaz|m>!newcalc 1/8 metric tons in lbs
11:09<linbot>Guspaz|m: (1/8) metric tons = 275.577828 pounds
11:09<Peng_>Apparently the US military is less strict about fitness than it used to be. :\
11:09<cdlu>Peng, ya, cannon fodder doesn't have to be in good shape.
11:09<spkitty>!newcalc 65 kgs in lbs
11:09<linbot>spkitty: 65 kilograms = 143.30047 pounds
11:09<Guspaz|m>peng_: That's because it's not a professional military. People enlist rather than apply.
11:10<Peng_>cdlu: In this case, it was someone applying for a desk job.
11:10<Guspaz|m>I used to be underweight. Now I'm 160lbs, so I'm fine.
11:10*cdlu was amused to find that Canadian police forces have a stricter fitness standard than the armed forces.
11:10*jess^ finds her rfid badge that lets her into work
11:10<Guspaz|m>We take policing seriously ;)
11:10<jess^>yay, it's not lost
11:10<cdlu>we do :)
11:11<cdlu>Peel Region has had more than 10,000 applicants already this year
11:11<cdlu>wouldn't want to be applying to be a cop at the moment!
11:11<Peng_>Which pays better?
11:11<cdlu>depends on your rank :)
11:11<cdlu>I'm sure a General in the armed forces makes more than a Special Constable in the police service
11:11<Peng_>Your average cannon fodder, as an example.
11:11<cdlu>cannon fodder pays poorly
11:12<cdlu>and they cut back your pay retroactive to the day you died
11:12<Guspaz|m>But if you're smart you probably won't end up as cannon fodder.
11:12<cdlu>"arrows cost money, the dead cost nothing" - longshanks (braveheart)
11:12<Guspaz|m>As in, they'd make you a tech or something.
11:12<erikh>someone's gotta run out there and get shot
11:13<erikh>besides, you guys are from canada
11:13<cdlu>I wonder if I still have "murphy's military laws" kicking around
11:13<erikh>all you guys do is crash planes and clean up horse crap from your "motorized transports"
11:13<erikh>sorry, had to be said
11:13<cdlu>gems like 'the buddy system is essential to your survival -- it gives the enemy someone else to shoot at'
11:13<cdlu>erikh, hey, at least we upgraded from air guns to 22s for our forces!
11:14<Guspaz|m>Those pellets packed a wollop, though...
11:14<cdlu>I had a friend who got all the way through basic without firing a shot
11:14<cdlu>bullets were too expensive :)
11:14<Guspaz|m>I'd imagine that simunition costs more than real rounds.
11:15<KingTarquin>Is there a way to view the entire zone file for a remote domain?
11:15<Peng_>cdlu: When was that?
11:16-!-HedgeMage [] has joined #linode
11:16<Peng_>KingTarquin: You can try an AXFR, but many domains don't allow it.
11:17<cdlu>Peng, my friend's basic?
11:17<cdlu>in the reserves
11:18-!-ph^ [~ph^@] has quit [Remote host closed the connection]
11:18<Peng_>I wonder if that's changed now, what with Afghanistan?
11:18<cdlu>no idea
11:18<cdlu>a friend of mine is going to be rotated into afghanistan next spring
11:19-!-Redgore2 [] has joined #linode
11:19*Guspaz|m wishes that the US had kept supporting Afghanistan rather than moving on to Iraq and leaving Canada with the slack :(
11:19*cdlu wishes we were fighting wars worth fighting
11:19<cdlu>why are we dying for another tinpot dictator?
11:20<Guspaz|m>Afghanistan never had a dictator.
11:20*cdlu gestures toward Karzai's slow seizure of absolute power
11:20-!-linville [] has quit [Ping timeout: 480 seconds]
11:20<cdlu>not a war worth fighting imo.
11:20<Guspaz|m>It was worth getting into to begin with.
11:21<Guspaz|m>As to if it's worth staying, dunno.
11:21<Guspaz|m>I've more or less ignored Canadian news for the past 6 months or so.
11:21<Guspaz|m>Including local :P
11:22<Guspaz|m>Ever since I stopped reading the newspaper for lack of time
11:22<cdlu>the news this morning noted that a Canadian soldier's chance of being killed in Afghanistan is now 1 in 62
11:22<linbot>New news from forums: Is Linode the right choice? in General Discussion <>
11:22<cdlu>the same chance as Russian soldiers had 28 years ago
11:22<amitz>cdlu: that doesn't sound good
11:22<cdlu>ya, it's pretty dire
11:23<JshWright>fun with statistics
11:23<cdlu>I wouldn't complain if it was worth fighting for
11:23<Guspaz|m>It's what happens when a country as small as Canada is tasked with occupying the entire southern portion of a country like Afghanistan.
11:24<amitz>JshWright: I know what you mean, to get that statistic, either Canada know something they should have told the people or there are disturbingly many incidents..
11:24<cdlu>ok, off to have some quintessential Canadian food
11:24*cdlu poutine &
11:24<amitz>or Lies, damn lies, and statistic.
11:25<amitz>sorry, wrong quote. "How to lie with statistic".
11:26*Guspaz|m is exagerating a bit. Canada only has to occupy the Kandahar province.
11:26-!-Redgore [] has quit [Ping timeout: 480 seconds]
11:26-!-Redgore2 is now known as Redgore
11:26<Peng_>Wikipedia says they plan to leave Kandahar either by or during 2011.
11:26<Guspaz|m>Meh, the plans to leave are decided by the whim of the party in power at the time.
11:27<Guspaz|m>It's a minority government. Whose to say that the Conservatives will be in power instead of the Liberals in a few months?
11:27<Guspaz|m>There's rumblings of an election soon.
11:27<Guspaz|m>Although I think the Bloc Quebecois propped up the conservatives recently to avoid the recent threat.
11:28-!-zack [] has joined #linode
11:28-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:1c6e:4d26:2fa:a654] has joined #linode
11:29<amitz>Guspaz|m: I get this impression that Canada's politic is very ..dynamic for a developed country?
11:29<jonny5>Cabbage is nearly ready on Farmville
11:30<Guspaz|m>amitz: Hardly, the kind of two-party system you have in the US is pretty rare in the developed world.
11:30-!-Frools [] has quit [Ping timeout: 480 seconds]
11:31<Guspaz|m>Also, fully private medical systems are, IIRC, unique to the US.
11:31<amitz>you mean, the 2 party system actually encourages intense dynamic instead?
11:31<JshWright>The US medical system is hardly fully private
11:32<Guspaz|m>JshWright: Apart from some mandatory state-imposed health insurance plans like MassHealth, or the US "MediCare" that is only for the old...
11:32<Guspaz|m>The US healthcare system is kind of a joke to the rest of the world.
11:32<Peng_>That still counts as "not fully private".
11:32<Peng_>Although I don't think anybody cares?
11:33<laser`>Oh, I just thought that Medicare was just a State-run insurance system?
11:33<laser`>the actual medical system is fully private?
11:33<laser`>There are state run hospitals too?
11:33<Guspaz|m>Are there? I don't know.
11:33<JshWright>laser`: yes
11:33<laser`>Learn something every day :)
11:33<Peng_>I haven't been paying attention; do the "Obama is a socialist!" people want to shut down Medicare too?
11:34<JshWright>Most state university systems have "teaching hospitals"
11:34*Guspaz|m is very happy with L'Assurance Maladie.
11:34*laser` <3 the NHS :P
11:34<laser`>I find this death list system horrendous :S
11:34<Guspaz|m>peng_: No, those people are busy saying the Government should keep their hands off Medicare, in a strange oxymoronic fashion, not realizing that the government runs medicare.
11:35<Peng_>Guspaz|m: Ah.
11:35<laser`>I can't believe some states (I remember reading about Flordia, unsure about others) actually have priority lists for cancer patients to get treatment in the case of a swine flu outbreak
11:35<Jonathan1>I don't understand why people would want to keep the healthcare insurance system in place, it seems very strange to me, stockholm syndrome
11:35<JshWright>laser`: why not?
11:36<Guspaz|m>Because here in Quebec, every single citizen can get a free swine flu shot if they want it?
11:36<Guspaz|m>The federal government ordered, IIRC, 52 million shots.
11:36<Guspaz|m>Which for some reason is 50% more than our population.
11:36<Peng_>The federal government of Canada?
11:36<JshWright>Guspaz|m: did they get them all?
11:36<Guspaz|m>JshWright: I presume so.
11:36<laser`>The UK government ordered 80 M, I think
11:36<laser`>Only 30 so far
11:36<Peng_>Well give some of them to us in the U.S.; I still haven't gotten a damn shot.
11:36<laser`>(30 M)
11:37<Battousai>sometimes you need a double dose of fun
11:37<Peng_>(It's scheduled for this week, though.)
11:38-!-Frools [] has joined #linode
11:38<JshWright>Guspaz|m: the issues in the US haven't been insurance related, they've been supply related
11:38<Guspaz|m>Apparently they're rolling it out on a priority-based system here too. Montreal got 160 thousand shots and those were given to highest priority. Next will be high risk people, although they'll vaccinate anybody who shows up.
11:39<Guspaz|m>So looks like NOvember 6th is when that happens.
11:39<Guspaz|m>JshWright: I know, but the US pays way more per-capita for health care and gets similar quality service to other countries that have universal healthcare.
11:40<Guspaz|m>I shouldn't care, though, since it doesn't affect me :P Unless I ever take a job in the US.
11:40<laser`>I thought a study showed that the quality was actually considerably lower?
11:40<JshWright>Guspaz|m: I'm not contesting that, I'm just pointing out places where you are mischaracterizing issues
11:40<Guspaz|m>JshWright: I wasn't mischaracterizing, I just randomly changed the topic.
11:41<Guspaz|m>I do that a lot.
11:41<Guspaz|m>laser`: I don't know that they do. Certainly it's no higher. But they *do* pay way more than us.
11:41<Guspaz|m>Because their healthcare providers are actually trying to make a profit of all things.
11:42<Guspaz|m>And their insurance companies are actually allowed to deny coverage and claims!
11:42<Guspaz|m>Kind of defeats the purpose of insurance.
11:42<laser`>"Domestic abuse" can apparently be a pre-existing condition :S
11:42<Guspaz|m>Yeah, I heard about a girl who got refused coverage because she wasn't eating enough and her parents took her to an eating coach.
11:43<jonny5>Can being denied coverage be grounds for denying coverage?
11:43<Guspaz|m>jonny5: It seems that they'll take any excuse.
11:43<JshWright>jonny5: in a sense, yes
11:43-!-orudie [] has quit [Quit: Leaving]
11:43-!-orudie [] has joined #linode
11:43<laser`>I think car insurance in the UK works like that
11:43<laser`>I know I normally get asked "Have you ever been denied coverage in the past?"
11:44<Guspaz|m>I just don't get it how they're allowed to deny coverage or deny claims. I mean, it boggles the mind that they're allowed to tell somebody they won't cover them.
11:44<laser`>They're private companies, they can cover whoever they want
11:44<Guspaz|m>That's what regulation is for.
11:44-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:1c6e:4d26:2fa:a654] has quit [Quit: When two people dream the same dream, it ceases to be an illusion. KVIrc 3.4.2 Shiny]
11:44-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:1c6e:4d26:2fa:a654] has joined #linode
11:45<Guspaz|m>Confuses the heck out of me. I'm not sure if our insurers can deny coverage, but medicare certainly can't, and everybody gets that.
11:47<Guspaz|m>I'm not sure if my health insurer (Empire Life) denies coverage. I've got lots of pre-existing conditions, but it was automatic when I got my job.
11:47<Guspaz|m>Awesomely, they provide supplemental coverage in the US (medicare only covers equivalent cost, and since US costs are insane, it doesn't cover all US costs)
11:48-!-AndrewLuecke [~kvirc@2002:7a6b:8d21:0:1c6e:4d26:2fa:a654] has quit []
11:51-!-A-KO [] has quit [Remote host closed the connection]
11:51-!-A-KO [] has joined #linode
12:02-!-powerrock [] has joined #linode
12:02-!-nb [] has quit [Ping timeout: 480 seconds]
12:02-!-MsMimi [] has joined #linode
12:03<MsMimi>Do you think it is better to buy a linode 2880 or split my servers and buy the linode 1440?
12:03<JshWright>MsMimi: what are you bottlenecking on?
12:04<JshWright>why do you need to upgrade?
12:04<MsMimi>cpu usage
12:04-!-elhippo [] has joined #linode
12:04<JshWright>what services are you running?
12:04<MsMimi>its over 150%
12:04<MsMimi>linode 1440
12:04<JshWright>no, what is your linode doing?
12:05<MsMimi>the cpu usage is so high
12:05<MsMimi>it gets to more than 200%
12:05<JshWright>I understand that...
12:05<JshWright>what tasks are you performing with your linode?
12:05<MsMimi>is the cause
12:05<MsMimi>i run a forum
12:06<Guspaz|m>You're using 150-200% out of 400%? Why do you need an upgrade?
12:06<Guspaz|m>You're using less than half your available CPU time.
12:06<MsMimi>oh really?
12:06<MsMimi>so 400% is max?
12:06<Guspaz|m>Four virtual cores, so yes.
12:06<spkitty>quad core
12:07-!-nb [] has joined #linode
12:07<JshWright>You could probably tweak MySQL to drop that usage down a bit...
12:07<MsMimi>I don't need to split yet
12:07<Guspaz|m>Well, not quad core, no.
12:07<MsMimi>i tweaked it already
12:07<MsMimi>its just that the search function
12:07<MsMimi>of the forum
12:07<spkitty>big forum then?
12:07<MsMimi>is killing my server
12:07<MsMimi>not that big
12:07<spkitty>usually the biggest split off search to a separate server
12:07<Guspaz|m>Linode hosts use two quad-core CPUs (eight total), the Xen virtual cores are not guaranteed to correspond to any physical CPU cores. They're virtual.
12:07<JshWright>rate limit the search a bit?
12:07<MsMimi>only 2000-3000 visitors
12:07<nb>spamassassin and clamav are killing my server
12:07*nb is upgrading to 720 linode
12:08<MsMimi>well JshWright
12:08<Guspaz|m>MsMimi: So, you've given MySQL a big enough cache, then?
12:08<powerrock>when I login, i get a message
12:08<powerrock>There is no screen to be attached matching linode16310.
12:08<powerrock>Your Linode isn't running, or another console session is already active.
12:08<MsMimi>that is the key to my success
12:08<MsMimi>tthe search function
12:08<MsMimi>if i limit it, it would bring down revenue and pageviews
12:09<Guspaz|m>Do you have any visible signs of performance problems?
12:09<Guspaz|m>Or are endusers happy with the performance?
12:09<MsMimi>well not really gus
12:09<MsMimi>its pretty decent
12:10<MsMimi>Though maybe if I upgrade
12:10<MsMimi>it would bring in more traffic
12:10<MsMimi>i was just asking
12:10<MsMimi>which is better
12:10<Guspaz|m>Well, if performance is fine now, I doubt an upgrade would help. Performance would go from fine to fine :P
12:10<MsMimi>ah thanks guys
12:10<powerrock>when I go into web dashboard, it says linode is up and running
12:10<MsMimi>your really helpful
12:10<Guspaz|m>If you're really high traffic, it might make sense to get a second linode and offload database stuff to that.
12:10<bliblok>Splitting is prabably better, since CPU is your bottleneck.
12:10<Jonathan1>MsMimi: You might want to look at something like pshinx
12:10<Guspaz|m>And access the database over the private network to avoid bandwidth charges.
12:11<powerrock>i see the site i hosted is up and running
12:11<jonny5>I've upgraded and performance feels worse, although need to put into figures or its just superstition
12:11<powerrock>what do I do?
12:11<Guspaz|m>powerrock: You might want to open a ticket with Linode, this is just community support.
12:11<powerrock>ok will do that
12:11<MsMimi>thanks jonathan
12:11<Guspaz|m>Sounds like a bit more than we can handle
12:11<MsMimi>i'll look into it
12:11<powerrock>i want to make sure that I am not doing something stupid :-)
12:12<Jonathan1>Mysql search facility gets killed on sites with lots of updates
12:12<Jonathan1>All the stuff mysql uses to short cut searches, is disabled everytime an update is made
12:12<quanin>powerrock: if you're inside lish now, try issuing a "kill" command, then reattach to it.
12:13<MsMimi>oh john we already do have full text searching
12:13<quanin>just in case your screen session's just dead from a previous connection that didn't time out.
12:13<Jonathan1>MsMimi - yes, but mysql full text searching isn't good in a heavy update environment
12:13<MsMimi>we get like 10 request per second
12:13<MsMimi>oh i c
12:13<Jonathan1>MsMimi: I assume you are just adding full text index to the existing tables?
12:13<MsMimi>yes Jon
12:14<Jonathan1>MsMimi: Could you deal with a deal in content being added, to it being available in search?
12:14<Jonathan1>sorry, deal with a 'delay'
12:14<MsMimi>I'll look into it jon
12:14<MsMimi>i am using vbulletin software
12:14<Guspaz|m>jonathan1: You're thinking, search in a MEMORY table that's periodically updated?
12:14<MsMimi>if that helps
12:15-!-kassah [] has joined #linode
12:15<nb>what do you think would be better? running everything on one linode 720 or having 2 360's and run clamd.scan and spamassassin on the 2nd?
12:15<powerrock>quanin, thanks, i tried that and it didn't fix the issue
12:16<Guspaz|m>nb: What are you limited on? RAM? Then one 720 should be fine.
12:16<powerrock>i will raise a ticket
12:16<nb>Guspaz|m, yeah, ram
12:16<Jonathan1>Guspaz|m: MySQL memory tables have very few advantages over disk based in my experience, but I was thinking of something along those lines, even having new content go to a seperate location initially, and only being bulked in hourly. This would improve the query cache etc performance a lot.
12:16<quanin>powerrock: try rebooting it? either from the web or from lish. if from the web, keep an eye on lish for job completed notifications.
12:16<nb>Guspaz|m, yeah, i have a 540 now, and tend to be running into swap quite a bit
12:16<jonny5>My 2 pennies, have you got APC or an opcode cache installed?
12:17<Jonathan1>MsMimi: I would maybe google and see what known plugins/solutions other users of the software use to power search
12:18<quanin>powerrock: failing that, yeah, definitely give a support ticket a try. but it sounds to me like lish just doesn't have the PID or whatever it needs to determine your linode's running.
12:19<Guspaz|m>nb: Then throwing more RAM at the problem should solve your problems. Doesn't sound like you need the extra headache of administrating two systems.
12:19<nb>Guspaz|m, true
12:20<Guspaz|m>I think we've got SpamAssasin running on our 540 at Otakuthon, and it's a huge memory hog.
12:21<MsMimi>thanks guys
12:21-!-nb_ [] has joined #linode
12:21<MsMimi>I'll do that jon
12:21<MsMimi>That would probably help me out a lot
12:21-!-MsMimi [] has quit [Remote host closed the connection]
12:22-!-kelvinq [] has joined #linode
12:22<jess^>mmm load averages
12:22<Guspaz|m>I was going to suggest to her that vBulletin apparently has some sort of built-in data caching system called "datastore caching"
12:23<nb>Guspaz|m, yeah, i changed it from 5 to 3 children, each take about 5% ram according to top, and clamd takes like 13%
12:23<nb>although clamd is just one process at least
12:23-!-MsMimi [] has joined #linode
12:23<MsMimi>got dc
12:23<jonny5>I was going to suggest APC cache if its in PHP... can sometimes half CPU usage
12:23<Guspaz|m>MsMimi: vBulletin apparently has built-in caching called "datastore caching", are you using that?
12:24<MsMimi>guspaz yes I am
12:24<MsMimi>I have fully optimized my forum
12:24<MsMimi>to the max
12:24<MsMimi>at least I think
12:24<Guspaz|m>Also using a PHP object cache? APC or eAccelerator or something?
12:24<jonny5>MsMimi: Have you installed munin or something similar to get graphs? They can help analyse the problem
12:25<MsMimi>yip using all the cache methods
12:25<Guspaz|m>If all else fails, you can move the database to a dedicated linode and use the private network to communicate between them, which is free (no bandwidth cost, high speed).
12:25<MsMimi>yes jonny its the mysql for sure
12:25-!-bogdanbiv [~quassel@] has quit [Ping timeout: 480 seconds]
12:25<MsMimi>dedicated linode?
12:26<jonny5>Dedicated database linode sounds like a good plan
12:26<Guspaz|m>Yeah, a linode just for the database server. So long as it's in the same datacenter, you've got the private network between them.
12:26<MsMimi>ah sweet
12:26<MsMimi>I had another question
12:26<jonny5>You could even get two of them and replicate between them
12:26<Jonathan1>You guys know more about linode than me, but if you only have 1 linode as a server - why add 1 for the db? Wouldn't it be better to just scale the existing linode, add the ram to mysql and have no network latency (however small)?
12:26<MsMimi>why should i stick with vps if i can get dedicated for the same price?
12:27<bliblok>Because linode rocks.
12:27<MsMimi>well i agree linode rocks
12:27<Guspaz|m>jonathan1: Increasing the size of a Linode doesn't necessarily give you much more CPU power.
12:27<MsMimi>i have not had a downtime yet
12:27<Battousai>free reboots!
12:27<Guspaz|m>A bigger linode *guarantees* you a bigger share, but if your host isn't heavily utilized anyhow, you'd get the same performance.
12:27<MsMimi>just asking why linode vps is better than say softlayer dedicated?
12:27<Jonathan1>Guspaz - I thought that was the whole point, larger linodes have less contention = more cpu share
12:28<Guspaz|m>You get a bigger share only if there is contention. If the host has idle time, you'll get it if you need it.
12:28<MsMimi>ah thanks
12:28<Jonathan1>Sorry, I still don't see the benefit in going to 1 frontend / 1 database linode
12:28<MsMimi>I guess i'll stick with linode until they kick me out
12:28<Jonathan1>It just seems to add network for no reason
12:28<Guspaz|m>jonathan1: If your linode is maxing out the CPU, there isn't much point to upgrading it.
12:29<Guspaz|m>MsMimi: It's a lot easier to deploy/manage linodes than a dedicated server, and you get a lot more control and flexibility.
12:29<Jonathan1>How can that be? If my linode 360 is maxing CPU, going to 1440 wont help?
12:29<Jonathan1>IF thats true, then somethings seriously wrong
12:30-!-kassah [] has quit [Ping timeout: 480 seconds]
12:30<Guspaz|m>jonathan1: Your linode can use up to half of a host's available CPU time if it's available. A 360 and 1440 both have the same theoretical maximum performance (half the host). The 1440 just guarantees a bigger minimum share.
12:30<Guspaz|m>But that max burst is still the same.
12:30<straterra>It'll only help if your 360 is fighting for CPU cycles with other guests
12:30<Jonathan1>So you do get more resources (assuming you always get min share) on a bigger linode, which makes sense
12:30<Jonathan1>Simple plain scaling
12:30<Jonathan1>Assuming worst case of min share.
12:31<MsMimi>ah thanks gus
12:31<Guspaz|m>Only if the host has CPU contention, which seems to be uncommon at Linode. The hosts are usually not overloaded.
12:31-!-eric_g [~chatzilla@] has joined #linode
12:31<MsMimi>is the best place to setup linode in New Jersey if your site is international?
12:31<Guspaz|m>MsMimi: That seems to be the consensus.
12:31<nb>yeah, most all of the hosts i have been on have been host load: idle
12:31<Guspaz|m>MsMimi: But that depends on where your visitors are mostly from.
12:31<Guspaz|m>nb: The host load is only for the dom0, not other linodes.
12:32<MsMimi>i see
12:32<eric_g>Anyone from management around?
12:32<Jonathan1>Yes, but lets assume the worst, then the basis of bigger linode = more CPU is true, what you might get for grattis on a smaller, is a hope
12:32<nb>Guspaz|m, oh so that doesn't tell how much is actually in use by other linodes?
12:32<Guspaz|m>nb: AFAIK, no.
12:32<MsMimi>jon won't it be more cost effective if you had 2 servers rather than one?
12:33<Jonathan1>MsMimi - if 1 server is the web server, and 1 server is the database... no. Having 1 larger server, running both would be more efficient.
12:33<Guspaz|m>jonathan1: Yes. My point is since linode hosts tend to be underutilized, if you're already maxing out half the host CPU, then upgrading, you'll still be maxing out half the host CPU. No benefit, then.
12:33<Jonathan1>The question comes when you need multiple web servers, then having a seperate database server starts to make sense.
12:33<MsMimi>ah i c
12:33<Guspaz|m>Also, disk IO is something that tends to be really hard to schedule, so it's incredibly difficult to fairly split disk IO.
12:33-!-kg4peq [] has joined #linode
12:34-!-jonny5 [] has quit [Quit: laters]
12:34<Guspaz|m>If you're RAM limited, upgrading the Linode makes sense. If you're IO or CPU limited, getting a second Linode makes more sense.
12:34<kg4peq>Is Atlanta having issues? Server is unresponsive, Console not loading.
12:34<Jonathan1>Guspaz|m: assuming you are bursting all the time, if one is within ones limits and simply wants to gurantee that those limits will surfice, then upgrading makes perfect sense
12:34<straterra>kg4peq: yes
12:34<Guspaz|m>kg4peq: They're having network issues in Atlanta:
12:34-!-darkbeholder [] has quit [Remote host closed the connection]
12:35<Guspaz|m>jonathan1: Yeah, and if you're maxing out the burst all the time and you want more, you need another linode.
12:35<Jonathan1>that makes no sense
12:35<Guspaz|m>Of course it does.
12:35<jess^>linode rocks
12:35-!-daMaestro [] has joined #linode
12:35<Jonathan1>no wait, it does
12:35<Jonathan1>Ok yes, it does
12:35<Jonathan1>However, your assuming it's maxing burst
12:35<Jonathan1>I'm talking about guranteed resources
12:35<@jed>^ hah
12:35<Guspaz|m>Jonathan1: Yeah, I am, but it seems to be the norm at Linode, and it's really difficult to tell if you are.
12:35<Jonathan1>and guranteed resources scales with size
12:35<jess^>jed: hi!
12:36-!-darkbeholder [] has joined #linode
12:36<Guspaz|m>jonathan1: The only way to ensure access to more burst capacity is another linode. You get both the minimum guarantee, and extra burst.
12:36<Jonathan1>Sure, you were starting to depress me, my plans of scaling at linode were looking like poo.
12:36<Jonathan1>I don't care about burst
12:36<Jonathan1>I'm happy to live within guranteed
12:37-!-linville [] has joined #linode
12:37<Guspaz|m>Yeah, the problem is, a customer on a 360 maxing out his CPU who upgrades to a 2880 might say "How come I'm still maxing out the CPU?"
12:37<Guspaz|m>And people have complained before.
12:38<Guspaz|m>Because they were getting max burst both before and after the upgrade.
12:38<Jonathan1>I understand, but maxing out is relative, the customer just isn't understanding that the cpu load still exceeds what they have available
12:38<Jonathan1>at this point you need to vertically scale
12:38<Jonathan1>I always mix those words up
12:38<Guspaz|m>Yeah, I was about to say.
12:38-!-eric_g [~chatzilla@] has left #linode []
12:39<Guspaz|m>xkcd's new design is awesome.
12:39<Guspaz|m>I wonder if any of my old geocities sites still exist. If they do, I haven't known their URLs for like a decade.
12:39<jess^>geocities closes today
12:39<Guspaz|m>I mean before it closes :P
12:39<Jonathan1>Thats fine, I've written my basic app to start with 2 linodes, the database is already in 2 shards. Initially I would simply expand horizontally for the database and vertically for the web servers, if it does enough traffic to need that much, I can worry about where to go next.
12:40<jess^>mmmm database shards.
12:40-!-Frools [] has quit [Ping timeout: 480 seconds]
12:40<Jonathan1>Todays xkcd was brilliants
12:40-!-nb [] has quit [Quit: Leaving]
12:40-!-nb_ is now known as nb
12:40<Guspaz|m>jonathan1: But we've all been dancing around disk IO. That's something that also doesn't scale terribly well. It tends to be easier to scale horizontally there.
12:41<amitz>This is probably the 4th time someone mentione about the coolness of the newest xkcd design.
12:41*amitz go see.
12:41<Jonathan1>Sure, but lets face it I'm doing it on vps, I'm doing it on a budget that wouldn't cover my ciggarets in the same time period, I can live with imperfections :)
12:42-!-MsMimi [] has quit [Remote host closed the connection]
12:42<Guspaz|m>A vps-based solution *can* be high performance. There are reasons other than limited budgets to go VPS. There are so many advantages to a well-done VPS (like at Linode).
12:42<Jonathan1>if disk IO becomes a huge issue, I'll throw everything into ram drives/memcached
12:43<Guspaz|m>Scaling horizontally rapidly, fast recovery from hardware failures, being able to clone nodes easily and move disk images around, easy failover solutions...
12:43<Jonathan1>Guspaz|m: I'm not nocking it, all I'm saying is, if I'd won the lotto, I'd probably do it different :)
12:44-!-darkbeholder [] has quit [Read error: Connection reset by peer]
12:44<Guspaz|m>If I won the lotto, I'd probably just scale horizontally and vertically at Linode. Except I no longer do any "professional" hosting (stuff that makes money).
12:44<Jonathan1>The easy failover is my primary reason for being here
12:44<Guspaz|m>I've been down the dedicated server route. Lindode is a breath of fresh air compared to that.
12:45-!-Bohemian [] has joined #linode
12:45<Guspaz|m>Only time I'd consider different is if I needed to start thinking about coloing.
12:45<Jonathan1>I've had very good and very bad dedicated experiences
12:45-!-atambo [~atambo@ATAMB0.RES.CMU.EDU] has joined #linode
12:45<Jonathan1>The good ones cost lots
12:45-!-darkbeholder [] has joined #linode
12:46<Jonathan1>I am very happy with linode, should it come to a point, which is unlikely as my idea will probably flop, where I need to scale like facebook, I will welcome the problem :)
12:46<Guspaz|m>Mine wasn't bad. I signed up at ThePlanet back when their support didn't suck. I was originally a ServerMatrix customer, from 2003.
12:47<Guspaz|m>Back then I used RHEL3, which couldn't upgrade between versions like Debian/Ubuntu can.
12:47<Jonathan1>Every unmanaged dedicated sever I've used, was a problem, mainly because of my lack of knowledge
12:47<Guspaz|m>And RHEL3 sucked.
12:47<Guspaz|m>Linodes are unmanaged though...
12:47<Jonathan1>I used dedicated servers from for a former contract, and they were brilliant, totally managed
12:47<Jonathan1>Yes, they are
12:48<tarpman>cdlu: I agree with whatever you just said.
12:48-!-Bohemian [] has quit []
12:48-!-Deckert [] has joined #linode
12:48-!-Frools [] has joined #linode
12:48<Jonathan1>My lack of knowledge isn't quite as painful at linode tho
12:48<Jonathan1>for all the reasons you stated
12:48<Jonathan1>and the great helpful irc guys :D
12:49<Guspaz|m>Yeah, it's nice that Linode has an active forum community like ThePlanet did, and also IRC.
12:52<Jonathan1>You've got me thinking about how to move now, it probably is better to add more nodes than upgrade them
12:53<Guspaz|m>Well, the same questions I asked the other people before, where are you bottlenecked?
12:53<Jonathan1>I'm not, it's just plan for scale, the first bottleneck will be ram.
12:53-!-ribtib [] has joined #linode
12:55<ribtib>just a quick question. Adding a second domain to my single linode, do I add a new dnszone and that's it (I know how to handle the virtual hosts stuff on the server later, I'm just not sure on the DNS settings)
12:55<Guspaz|m>RAM is quite well handled by vertical scaling.
12:55-!-hercynium [] has joined #linode
12:55<Jonathan1>I guess the first move will be to seperate the frontends from the db - go to 4 small nodes, then can go verticle from there
12:56<Jonathan1>ribtib: yep, 1 zone per domain
12:57<ribtib>Jonathan: thanks much. anything else around the dns setting I need to think of for my second domain or is it just a simple as that. just want to doublecheck if there's something obvious I'm missing.
12:57<Jonathan1>Nup it's pretty straight forward, I only did it yesterday myself, nothing tricky
12:58<ribtib>ok, thanks.
12:58-!-ribtib [] has quit []
12:58<Guspaz|m>The nice thing about databases is that, with replication, you can scale that horizontally as much as needed. Actually, you can do the same with application servers if you've got a reverse proxy frontend. I hear that Varnish is quite nice at that.
12:59<Jonathan1>I'm using haproxy to load balance/failover the web/app servers
12:59<Jonathan1>I'm working on it doing the database failover/load balancing
12:59<Jonathan1>having 2 db shared, both replicated
12:59<Jonathan1>sorry, shards not shared
13:02-!-dajhorn [] has joined #linode
13:02<Jonathan1>My main problem with the moment is php
13:03<Jonathan1>I didn't want to write the basic web parts in java, and I don't like ruby/python
13:03<Jonathan1>but php frameworks are so poor for anything that isn't bog standard 1 server 1 db
13:03-!-JoshM [] has joined #linode
13:05-!-JoshM [] has quit [Remote host closed the connection]
13:05<nard>what framework are you using?
13:05<@mikegrb>mmm cake
13:05<Jonathan1>I flipped a coin and went with zend, symfony/cake all have the same issues
13:06<Jonathan1>I think it's more down to the orm's than the frameworks, the orm's don't like sharding
13:06<@mikegrb>mmm cake
13:06<nard>ah. I've used cake myself. you really need to tune the queries, because sometimes their ORM is chatty and pulls alot of extra data or makes unneeded joins for the situation.
13:06<Jonathan1>I'm trying to get doctrine to accept that modelX is sometimes related to connection1, but sometimes connection2
13:07<Jonathan1>I'm blessed/cursed in that there are pretty much no joins in the app
13:07<Jonathan1>because of the shards, you can't join anything really
13:07<nard>what you'll need to do, is either do modulus to decide which queries goto what database, or you need to connect probably to a main database, to make a query, find out what database this query can run on, and run it again
13:08<Jonathan1>I'm going with a hash mechanism in the primary keys, not using numeric keys
13:09-!-kenichi [~kenichi@] has joined #linode
13:09<Jonathan1>the keys are essentially base 62 representations of sequential ints, but with some detail to define the shard it'll be in
13:09<Jonathan1>hopefully performance won't be too painful as they are still pretty small strings
13:09<nard>can't you use that data then, to tell your php framework where to look for data then?
13:10<Jonathan1>Yep, it's the actual mechanism of telling doctrine to do it thats a pain in the butt
13:10<Jonathan1>Doctrine relies on modelX always going to db1
13:11<Jonathan1>Have to hack it a bit and tell it to forget everytime the same model from a different shard is required
13:11<Jonathan1>It's just a whinge, it would be nice if the functionality was a bit more available, it's not impossible
13:12<nard>yea, thats i guess the challenge of going a non standard route =p. I've had to hack up cakephp alot to do some things out of the ordinary.
13:13<Jonathan1>I haven't worked in php in a few years, I thought the frameworks would be very different from how they started
13:13<Jonathan1>but they seem to have the same problems
13:13<nard>oh, an idea you can probably do, is make a modification to the orm in your framework to take in an extra value, and use that value to tell the orm which db to query. that way you can still use the orm, with minimal changes.
13:14-!-elfgoh [] has quit [Quit: Leaving]
13:15<Jonathan1>Basically what I'm doing at the moment is passing the key of whatever type of item, to a class that returns the db connection of the shard it's on, doctrine is then 'brainwashed' into forgetting previous locations and to assume all of those models are now in this shard
13:15<Jonathan1>the application makes sure it splits up queries to the shards so that nothing goes missing, and it works ok
13:15<Jonathan1>it just took a long time to work out how to brainwash doctrines internals
13:16-!-Bohemian [] has joined #linode
13:17-!-hfb [] has joined #linode
13:17-!-Bohemian [] has left #linode []
13:17-!-JM [] has quit [Read error: Connection reset by peer]
13:18-!-Bass10 [] has joined #linode
13:32-!-pharaun [] has joined #linode
13:33-!-zack_ [] has joined #linode
13:33-!-kg4peq [] has quit [Remote host closed the connection]
13:33-!-zack [] has quit [Read error: Connection reset by peer]
13:33-!-zack [] has joined #linode
13:40-!-kelvinq [] has quit [Quit: kelvinq]
13:41-!-zack_ [] has quit [Ping timeout: 480 seconds]
13:44-!-praetorian [] has quit [Remote host closed the connection]
13:45-!-praetorian [] has joined #linode
13:52<linbot>New news from forums: 2 Queries - Changing ssh port and ajax console in General Discussion <>
13:54<@caker># uname -a
13:54<@caker>Linux #1 SMP Mon Oct 26 17:41:12 UTC 2009 i686 GNU/Linux
13:54<@caker>can I get a witness?!
13:54-!-kelvinq [] has joined #linode
13:54<Battousai>no you may not
13:55<Guspaz|m>So we're finally moving to a non-ancient kernel. Harumph.
13:55<Guspaz|m>Nothing, I'm just being contraryminded.
13:55<Battousai>non-ancient kernels have been around quite a while
13:55<@caker>we've always had more recent kernels available
13:55<Guspaz|m>Not as stable :P
13:56<@caker>what makes you think this one will be marked stable?
13:56<Guspaz|m>A false assumption.
13:56<@caker>damn straight
13:56<Guspaz|m>Don't mind me, I'm annoyed at compile issues, I'd negate anything you said. You could say that the sky is blue and I'd argue.
13:56<Battousai>not at night it isn't!
13:57<Nivex>Sure it is! It's just so dark you can't see it.
13:57<JshWright><whisper>He's also Canadian</whisper>
13:57<Nivex>caker: are you making that "latest"?
13:57<@caker>no, but I'm considering making a new "Latest 2.6 paravirt"
13:57-!-zack [] has quit [Read error: Connection reset by peer]
13:58<Guspaz|m>caker: What's keeping newer than 2.6.18 from being marked stable anyhow?
13:58-!-zack_ [] has joined #linode
13:58<Battousai>we need a "Latest kernel32.exe"
13:58<@caker>Guspaz|m: the fact that thousands of VPSs don't have NTP installed
13:58<@caker>also, pvops has just recently become non-buggy
13:59<Guspaz|m>Battousai: Should be theoretically possible, but doesn't seem to be in Linode's plans.
13:59<Guspaz|m>caker: Oh, the time issues. I thought they were solved by now.
13:59<Nivex>If it's a matter of "Latest 2.6 paravirt" being blessed-just-make-sure-you-run-NTP then blog that sucker and let 'er rip :)
14:01<Nivex>for reference:
14:01-!-zack_ [] has quit [Read error: Connection reset by peer]
14:02-!-zack_ [] has joined #linode
14:05-!-sc0field [~rajiv@] has quit [Ping timeout: 480 seconds]
14:07-!-hercynium [] has quit [Ping timeout: 480 seconds]
14:08-!-reillyeon [] has quit [Quit: leaving]
14:08-!-reillyeon [] has joined #linode
14:09<Guspaz|m>So if pvops recently went stable, then presumably it will be possible to declare it as stable some time after the time issues are resolved.
14:11-!-Bohemian [~Bohemian@] has joined #linode
14:11-!-kelvinq [] has quit [Quit: kelvinq]
14:14<Battousai>never trust the customer to do the right thing
14:14<spkitty>pretty sure that's oneof the ten commandments Battousai
14:14<Battousai>invariably, some people will see ntpd, think they don't need it and rip it out. then they'll complain that the time is wrong
14:15-!-praetorian [] has quit [Remote host closed the connection]
14:15<JshWright>I hate it when a given script, typed verbatim into the interpreter works just fine, but fails silently when run from a file...
14:17-!-bogdanbiv [~quassel@] has joined #linode
14:17-!-praetorian [] has joined #linode
14:18<JshWright>though I suspect the issue actually lies (lays?) in the API I'm calling
14:18<Guspaz|m>Battousai: What's the underlying cause of the inaccurate time? Presumably that'll be fixed at some point?
14:19-!-Bohemian [~Bohemian@] has quit [Quit: Bohemian]
14:22-!-Bohemian [] has joined #linode
14:22-!-Bohemian [] has left #linode []
14:32<Peng_>Guspaz|m: The problem is that in pv_ops kernels, the clock is independent from the host's, so obviously the guests need NTP.
14:33-!-nard [] has quit [Quit: nard]
14:35<Guspaz|m>That implies that the guests are incapable of keeping accurate time without help...
14:35-!-ph_ [~ph^] has joined #linode
14:36<Guspaz|m>There would have to be a solution to this eventually, I mean, Linode can't stay on 2.6.18 forever. By 2093, I expect it will cause issues!
14:36-!-supine [] has quit [Quit: Leaving.]
14:37<Desph>Any linode coupon codes laying around?
14:37<JshWright>keeping proper time is all but impossible for a VM
14:37<Peng_>Keeping proper time is all but impossible for any computer.
14:38<JshWright>either it needs the host to update it regularly, or it needs to do it on its own
14:38<Peng_>Hell, my Linode is better at it than my PC at home, but nonetheless it isn't an atomic clock.
14:38<Peng_>Desph: Not usually.
14:38<Peng_>Desph: You do, of course, gte a discount if you pay for 1+ years.
14:39<Desph>Yeah I bet. I'm having trouble paying for -1 month right now haha
14:39<Guspaz|m>JshWright: Wouldn't adding ntp to the base images fix that for new linodes?
14:39<jess^>i'll never have problems paying for my linodes
14:39<JshWright>Guspaz|m: I reckon it would, yes
14:39<jess^>because they bill on the first and i get paid on the 30th :D
14:40-!-nard [] has joined #linode
14:40<Desph>Good for you >.>
14:40<Guspaz|m>You only get paid once a month o_O
14:41<Peng_>There's nothing _wrong_ with a pv_ops domU's clock, any moreso than any other Linux computer; it's just independent from the host's, requiring NTP to make up for the drift.
14:42<amitz>Guspaz|m: I pay my subordinates 5 times a month in average :-p
14:42<Peng_>If Linode _did_ start putting NTP on the images, I really hope they'd set up some NTP servers for people to use.
14:42<Guspaz|m>peng_: Then the problem will never go away, and it's not really a reason to consider pvops kernels unstable. Unless they'd intend to stick with 2.6.18 for all eternity.
14:43-!-cpg [] has joined #linode
14:44<Guspaz|m>Even on pvops machines? Then a reboot would eliminate drift. It would slightly mitigate the severity.
14:44<Guspaz|m>Of the issue.
14:44<@caker>SpaceHobo: that's correct on native xen kernels (2.6.18). On pv_op kernels there's no such connection
14:44<jess^>so would putting ntpdate in the cron.hourly :D
14:45-!-mortie [] has quit [Read error: Connection reset by peer]
14:45-!-Smark[Gone] is now known as Smark
14:45<Peng_>jess^: That's not funny; there are too many people who do things like that.
14:46<jess^>Peng_: how is that not funny?
14:46<Desph>Peng_: Whys that a bad thing? apple time works
14:46<jess^>the reason clocks on modern windows computers and apple computers DON'T drift is BECAUSE they timesync several times a day
14:47<jess^>vista uses a microsoft nptdate server
14:47<jess^>os X uses
14:47<jess^>so why can't linux users do the same with their systems?
14:47<Guspaz|m>ntpd corrects for drift in realtime, having the kernel adjust the rate at which time flows to compensate.
14:47<Desph>Because Peng_ said so
14:47<jess^>ntpdate is in my cron.daily
14:47<Peng_>Because Peng sucks at explaining things.
14:47<adj>ntpd also only corrects a fixed amount. if you are drifting faster than that it will never catch up
14:48<jess^>yeah but
14:48<jess^>i mean
14:48<Guspaz|m>Syncing daily means that every day, your clocks will all jump and your logs won't make sense.
14:48<Guspaz|m>adj: I believe that ntpd will make a larger jump if it has to.
14:48<Peng_>Because it's rude to run ntpdate at the exact same time; a lot of people do it, so NTP servers get _much_ more load at e.g. the top of the hour.
14:48<jess^>ok, BRAND NEW BOX, up for 7 days, and
14:48<adj>believe whatever you want =)
14:48<jess^>26 Oct 18:45:47 ntpdate[14799]: adjust time server offset -0.000563 sec
14:48<Peng_>adj: By default, NTP will step the clock if it's more than 128 ms off.
14:48<jess^>.000563 is a _tiny-assed_ number
14:49<Guspaz|m>jess^: Why would you do that when you can just "sudo aptitude install ntp" and your clocks will all stay synced?
14:49<Peng_>That too.
14:49<Peng_>Plus, is open access?
14:49<jess^>Guspaz|m: why would i do that when ntpdate is already there?
14:49<jess^>Peng_: yes
14:49<jess^>as far as i know, it is
14:49<straterra>I'd rather use ntpd too
14:50<Peng_>jess^: Because NTP is better?
14:50<jess^>and even if it isn't i own a mac so i have the rights to use it
14:50<jess^>Peng_: better hwo
14:50<jess^>better how
14:50<Peng_>jess^: And to avoid the load issue I mentioned above.
14:50<adj>jess^: if you use ntdp and tier your time servers and clients properly, you *should* never have to cron it
14:50<jess^>"what does it do, and how well does it do it"
14:50<jess^>when comparing two itesm
14:50<jess^>L is for Lysdexia
14:50<adj>jess^: and all you timing sensetive apps will work and your log servers will all have the right timestamps
14:50<Guspaz|m>jess^: Because it's the proper way to do things, it's better, corrects for drift, it's easier to set up, and doesn't hammer Apple's server at a certain point in time?
14:50<Desph>lol @ spell fail
14:51<Guspaz|m>You could manually muck with cron stuff, or you can just run one command and never worry about it again.
14:51<Guspaz|m>All the while getting better accuracy.
14:51<Desph>ima run aptitude install ntp on mah linode
14:51<Peng_>Even if you don't care about _your_ clock, the "hammering Apple's server" thing is rude.
14:51-!-powerrock [] has quit [Quit: powerrock]
14:51<amitz>Guspaz|m: ntpdate says that it's good for ong term? what to use instead?
14:52<Guspaz|m>amitz: Instead why? Huh?
14:52<adj>Peng_: so is setting your dns servers to and, but everyone does that too ;)
14:52<linbot>New news from forums: Remote Authentication for Mail Users in Email/SMTP Related Forum <>
14:52<jess^>hammering apple's server? excuse me? do you know how many fucking macs out there >already< sync with
14:52<straterra>Yes..but they run ntpd :P
14:52<jess^>also, how much data traffic are we TALKING about?
14:52<Guspaz|m>amitz: Just install ntpd ("sudo aptitude install ntp") and be done.
14:52<Guspaz|m>jess^: Those macs don't all hammer their server at the same TIME.
14:52<Guspaz|m>You're operating on a fixed schedule.
14:52<adj>jess^: a TREMENDOUS number of small packets, actually
14:53<adj>i'd love to know what arch the firewall in front of that time server is
14:53<Desph>The following NEW packages will be installed:
14:53<Desph> libcap1{a} ntp
14:53<Desph>The following packages will be REMOVED:
14:53<Desph> procmail{u} sensible-mda{u}
14:53<erikh>and if you're hurt about costing apple money, use
14:53<erikh>then you (well, if you live in the US) pay for it too.
14:53<jess^>i installed ntpd :P
14:53<amitz>Guspaz|m: oh, it's on ntpdate description. But ntp package has no such disclaimer so I'll use that instead.
14:54<Peng_>erikh: Nooo! Don't abuse stratum 1 servers! :(
14:54<erikh>Peng_: if they didn't want to be abused, they'd block me
14:54<Peng_>jess^: <3
14:54<erikh>and I pay taxes, goddamnit.
14:54<Desph>So whats more idel- ntp or ntpdate
14:55<erikh>ntpdate is better than telling your ntpd to correct large amounts of drift
14:55<erikh>ntpd is better for the typical use.
14:55<erikh>SpaceHobo: nice.
14:56<Guspaz|m>Ubuntu uses I believe.
14:56<erikh>yeah, i'll stick to the combo until I see it in action
14:56<erikh>it's tried and tested
14:56*adj is with erikh
14:56<adj>periodically ntpdate. use ntpd for small drift
14:56<Guspaz|m>ntpd has been tried and tested for practically forever.
14:56<adj>ntpd only does a max of 500 ppm drift
14:57<amitz>ntpdate: ntpdate is a simple NTP client that sets a system's clock to match the time obtained by communicating with one or more NTP servers. It is not sufficient, however, for maintaining an accurate clock in the long run. ntpdate by itself is useful for occasionally setting the time on machines that do not have full-time network access, such as laptops.
14:57<Guspaz|m>peng_: They must have upated it.
14:57<amitz>Guspaz|m: ^^
14:57<Guspaz|m>So, yeah, even the "tried and true" solution says you shouldn't be using it :P
14:57<Peng_>adj: Not many clocks are broken enough to drift >500 PPM.
14:58<adj>Peng_: look at the ven/vmware bugs from a few years ago. time keeping was awful (and still rears its ungly head on vm's occasionaly)
14:58<Peng_>adj: If NTP is doing fine, there's no reason to do a "periodic" ntpdate. Hell, you have to temporarily turn off NTP to do it; that's just pointless.
14:58<Peng_>adj: Ah. Good point.
14:58<amitz>adj: periodically as in maybe once a month?
14:59<adj>amitz: once a week maybe. tbh, its usually sometime i do manually. i've never had it become an issue except of virtual machines
14:59<adj>i also am having major failures today getting what i'm thinking into words as i type ;)
14:59<Guspaz|m>The question is, does ntp only do a big update once when it starts, or does it do it whenever the drift is too big?
14:59<adj>Guspaz|m: only when it starts
15:00<amitz>okay, ntp then. but set it to before installing
15:00<erikh>you pretty much can't manage a freebsd VM in virtualbox without ntpdate
15:01<erikh>especially if you suspend.
15:01-!-dvyjones [] has quit [Quit: Probably never]
15:01<Peng_>I doubt Apple intends for non-customers to use Of course, everybody's probably bought something from Apple at some point... :P
15:01-!-dvyjones [] has joined #linode
15:01<quanin>not me. :P
15:02<erikh>Peng_: if it really mattered, they would have used a proprietary protocol.
15:02<erikh>(hint: it probably doesn't)
15:02<adj>you could go buy a pair of gps receivers if you *really* need accurate timing
15:02<nard>yea, i dont think they mind as long as your not seriously abusing their servers. if anything they would block you or send an email to the people managing your servers letting them know
15:02<laser`>I just use tbh
15:02*rainman` runs some boxes with GPS antennas attached for timing
15:02<rainman`>fun stuff
15:02<nard>yea. i use
15:03<Guspaz|m>Does anybody have any idea why, after installing munin and munin-node, I have no graphs? (
15:03<adj>rainman`: sonet gear?
15:03<rainman`>adj, trimble antenna with converter to rs232, and custom freebsd driver
15:03<Guspaz|m>munin-node is running, the graphs are getting generated every 5 minutes, the symlinks to the plugins are there...
15:03<rainman`>adj, :)
15:03<adj>Guspaz|m: telnet localhost 4949 and verify you can run 'list'
15:03<adj>then telnet from the master munin server
15:04<adj>then try a config service and fetch service.
15:04<Guspaz|m>It's just one box, it's all local.
15:04<adj>if that all works, you have a permissions issue on the graph dir
15:04<nard>heya guys, i have a question about monitoring tools, and was wondering if you have any recommendations that are free. i've used cacti before. I need to monitor mysql cluster, web servers, memcache and load balancer. so even if it cant get the data directly, if it's able to run a script and grep the results, that would be good too.
15:04<adj>rainman`: fun fun =)
15:04<Guspaz|m>I telnet local and execute "list" and it outputs a blank line.
15:05<adj>Guspaz|m: restart the munin-node service
15:05<adj>and try again
15:05<adj>its not loading your plugins.
15:06<Guspaz|m>That fixed it.
15:06<Guspaz|m>But, I wonder why it didn't work after install.
15:06<Guspaz|m>Shouldn't have to manually restart it after installing.
15:06<adj>nard: use munin like Guspaz|m is =) it rocks until you get over 150-200 boxes
15:07<nard>ah k cool. i'm running like 7-8 linodes atm lol
15:08<jess^>linode+istatd + iphone = win
15:08<Yaakov>jess^: I set it up and it is OK, but I didn't find it as exciting as all that.
15:08<jess^>i think it's neat. :D
15:09<adj>Guspaz|m: and i *highly* recommend setting up the linux_diskstat_ plugin
15:09<Yaakov>jess^: It's OK.
15:09<adj>because iostat is more or less useless for trending disk usage
15:09<Yaakov>jess^: I wish it had more functionality, but it is helpful to get an idea that everything is OK.
15:10<Yaakov>And the phone part is neat.
15:10<jess^>Yaakov: it's in active development! suggest stuff!
15:10<Yaakov>I mean seeing the phone stats.
15:10<jess^>NOT HERE
15:11<Yaakov>jess^: I am confident that others will suggest the things I have in mind.
15:11<jess^>first they came for the jews, and i did not speak up, for i was not a jew
15:11<Guspaz|m>adj: I believe that's already included in munin-plugins-extra.
15:12<adj>Guspaz|m: for debian/ubuntu it is, iirc
15:12<Yaakov>I am a Jew, however, it has nothing to do with this app.
15:12<jess^>jewish people rock
15:12<jess^>if i end up dating one more, i'll have collected the whole set
15:12<Yaakov>Some of us do.
15:12<Guspaz|m>Hmm, if it's included, I don't see it, so it may not be enabled by default.
15:12<jess^>my current boyfriend is pagan. :s C-C-COMBO BREAKER
15:12<adj>Guspaz|m: i think its in ubuntu. check /usr/share/munin/plugins
15:13<jess^>i <3 this blouse.
15:13<jess^>it's comfortable.
15:13<Yaakov>My shirt isn't bad.
15:13<Guspaz|m>Nope, not seeing it :(
15:13<jess^>but i never understand why people make clothes with buttons that don't do anything
15:13<jess^>like they're just some sort of facade
15:14<quanin>they do it to piss us off.
15:14<adj>Guspaz|m: then i'd grab it =) if you ever grow out of your VM, you are going to want to know iops and latency numbers, not the dumb MB/s and block/sec that the default gives you
15:14<quanin>and because they know someone's sitting on some IRC channel wondering why.
15:15<Guspaz|m>Oh, I'll never outgrow it. It only has one high-traffic site, and I'm hosting that for free.
15:15<Guspaz|m>An ad-free high-traffic site.
15:15<jess^>SpaceHobo: who?
15:15<Guspaz|m>Anyhow, I need to grab some lunch. 'Tis after 3.
15:16<jess^>this weekend i think i'm going to make paella.
15:17<Guspaz|m>Apparently linux_diskstat was renamed diskstat_ and merged into munin's trunk.
15:17<Guspaz|m>So eventually we'll see it as part of the default install.
15:18-!-jforman [] has left #linode []
15:19-!-compwhizii [] has joined #linode
15:25<jess^>i'm a wiccan.
15:28<Yaakov>Also, on the lore of Steamboat Willy, the first Mickey Mouse cartoon.
15:32<Yaakov>Ford wasn't a very nice man.
15:32<Yaakov>Not like WALT DISNEY was was a SAINT.
15:33<spkitty>walt disney was a bastard
15:33<spkitty>he fucked over his own brother too
15:34<Peng_>MR. ROGERS
15:34<Yaakov>I am not so sure about that. Hitler had members of his inner circle who authored their own oddball pagan-teutonic-aryan mythology with a Wagner soundtrack.
15:34<Jonathan1>I hear mother Teresa could be a bit of a premadona
15:35<Yaakov>spkitty: He was a SAINT. A perfect soul.
15:35<JshWright>Speaking of people Hitler liked, I find great irony in the fact that two pieces of music played together perhaps more than any other... Wagner's "Bridal Chorus" and Mendelssohn's "Wedding March" were written by a anti-semite, and a Jew, respectively
15:42<adj> <-- beautiful
15:43<Guspaz|m>Not enough blinkenlights.
15:44<jess^>Alles touristen und non-technischen looken peepers! Das machinkontrol is nicht for gefengerpoken und mittengrabben. Oderwise is easy schnappen der springenverk, blowenfus, undpoppencorken mit spitzensparken. Der machine is diggen by experten only. Is nicht fur geverken by das dumpkopfen. Das rubber necken sightseenen keepen das cotton-picken hands in das pockets. So relaxen, und vatchen das blinkenlights.
15:45<adj>Guspaz|m: the switch sure blinks plenty
15:45<@caker>you rang?
15:45*Guspaz|m laments the startling reduction in blinkenlights in consumer-level switches.
15:46<Guspaz|m>I've even seen some switches go down to JUST ONE LED PER PORT.
15:46<jess^>green for link, blinken for act, orange for 10Base-T, red for fuct.
15:47<jess^>tri-color leds are easy to make :D
15:47<Guspaz|m>Not even that.
15:47<Guspaz|m>They just have the green for link and blinken for activity.
15:48<Guspaz|m>It's dispicable.
15:48<JshWright>Like the ProCurve's where you have to toggle through the different settings for information the LED is representing
15:48<Yaakov>caker: I felt that was very important.
15:49<@caker>thanks :)
15:49<Yaakov>I am always looking out for your best interest.
15:50<Peng_>Guspaz|m: My consumer-grade router just has that, but I've always found it nicely blinky.
15:50<Peng_>I guess I have nothing better to compare it to.
15:50-!-darkbeholder [] has quit [Remote host closed the connection]
15:50<Peng_>(To be clear, it has a pair of lights for each port.)
15:50-!-Keith-BlindUser [~blah@] has joined #linode
15:51<linbot>Keith-BlindUser: [spf] none: does not designate permitted sender hosts (spf record: )
15:51<Keith-BlindUser>!dig txt
15:51<linbot>Keith-BlindUser: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 14420 IN SOA 2009102632 7200 7200 1209600 14440
15:51<@caker>Keith-BlindUser: you know that stuff works in private message, right?
15:51<@caker>rather than spamming almost 300 people
15:52<Guspaz|m>!spam caker
15:52<linbot>SelfishMan: FDIC has officially named your bank a failed bank
15:53<Keith-BlindUser>caker: Sorry about that. I'll make sure that stuff gets sent in PM for now on.
15:53<@caker>Keith-BlindUser: cheers
15:55<Guspaz|m>You just need to add "/msg caker " before each command ;)
15:56<Peng_>Oh, that's mean.
15:56<Guspaz|m>Damn you, Intel. I spend $700 on a hard disk and you refuse to release firmware for it that supports TRIM.
15:57<Peng_>(Note: Mean in a funny way.)
15:59<Keith-BlindUser>caker: How did Linode get to be so amazing? It just seems to run perfectly! Well, maybe not perfectly..but still.
15:59<Guspaz|m>Regular virgin sacrifices.
15:59<Jonathan1>and how do you get those whites so white?
16:00<Guspaz|m>It makes your whites whiter, your brights brighter.
16:00-!-johndbritton [] has quit [Quit: johndbritton]
16:00<aquillar>How do you get those chipotle stains out!
16:01<Jonathan1>I had to google, but chipotle sounds good
16:01<Guspaz|m>Hi, Billy Mays here for Linode. Want to put the "fun" back in "functional"? Linode has the management interface to help YOU have a great time instead of slaving away on your server!
16:01<Jonathan1>but wait, theres more
16:02<Jonathan1>free case knifes, they dont dent and they dont blunt, free, 19.95 delivery
16:02<Guspaz|m>Billy Mays could sell anything to anybody.
16:02<Jonathan1>no idea who he is :-?
16:02<Guspaz|m>Just pay separate handling.
16:02<Jonathan1>We started getting usa infomercials last year or so
16:02<Jonathan1>they are freaky
16:02<Guspaz|m>He's the most awesome pitchman in the history of everything. Also he had a Discovery channel show that completed the first season shortly before his death.
16:02<Jonathan1>the same thing repeated for 6 hours
16:03<Guspaz|m>He became such a big celebrity that they had to get a private jet because every time they tried to go to the airport he'd get mobbed for autographs.
16:03<Jonathan1>lol cool
16:04<Guspaz|m>He was loud, always wore the same thing, very distinctive.
16:05-!-atambo1 [~atambo@ATAMB0.RES.CMU.EDU] has joined #linode
16:05<Keith-BlindUser>Linode is really cool!
16:06-!-atambo [~atambo@ATAMB0.RES.CMU.EDU] has quit [Ping timeout: 480 seconds]
16:06<Keith-BlindUser>BTW, I wrote a review of Linode for the blind user. It can be found at:
16:06<@caker>ok, two new kernels released, along with a new "Latest 2.6 Paravirt" pointer (for 32 and 64 bits)
16:06-!-jas4711 [] has quit [Quit: Ex-Chat]
16:07<Keith-BlindUser>TWo new kernels?
16:07<Keith-BlindUser>er two
16:07-!-kassah [] has joined #linode
16:07-!-azaghal [] has quit [Ping timeout: 480 seconds]
16:07<bd_>caker: I seem to remember some hosts having issues with recent kernels a while back; has that been fixed?
16:08<bd_>eg newark126
16:08<@caker>it's because of the new paravirt spinlocks implementation which some of our older hosts don't support. We decided to leave it compiled in since it's neat
16:08<bd_>what .config entry is that?
16:08<@caker>newark126 is not affected
16:09<bd_>hmm, I could've sworn I was having booting issues right after migrating to newark126 that required me to go back to .18
16:09<bd_>I suppose I'll have to give it another try one of these days
16:10<Keith-BlindUser>Is Dalass effected, and is it necessary to update the kernels?
16:10*Keith-BlindUser is on 2.6.30.whatever
16:10-!-azaghal [] has joined #linode
16:11<Peng_>Keith-BlindUser: 1.) You'll have to be more specific than "Dallas", 2.) It's
16:11<Peng_>, to be specific. On 32-bit, anyway.
16:11<Keith-BlindUser>Yeah that's what I'm running
16:11<Peng_>caker: What's cool about the new kernels?
16:11<Keith-BlindUser>I thought that Caker just said that thre was new kernel releases though?
16:12<@jed>correct, we've released two new kernels
16:13-!-awnstudio [] has joined #linode
16:13<JshWright>jed: is this the big surprise?
16:13<@caker>also, for whaever reason, this kernel doesn't always report ~1.5% cpu when idle, like some old pv_ops (and host combinations) did
16:13<@jed>JshWright: no :^)
16:13*Peng_ is never idle anyway.
16:14<Keith-BlindUser>caker: what does the kernel do instaead?
16:14<Peng_>That's a long list of changes.
16:14<Peng_>caker: If you don't mind, what new changes are interesting for Linode?
16:15<Keith-BlindUser>Got to run!
16:15<Keith-BlindUser>I'll check the logs back later, as I know that logs are generated online!
16:15-!-Keith-BlindUser [~blah@] has quit [Quit: Be back later!]
16:15<@caker>Peng_: I turned off CONFIG_SYSFS_DEPRECATED_V2 is about it
16:15<@caker>otherwise it's just an updated kernel in all its glory
16:16<Peng_>Well, I'm in no hurry to leave
16:20<Peng_>What does CONFIG_PARAVIRT_SPINLOCKS=y mean?
16:21<@caker>virtualization aware spinlocks in the guest
16:21<Peng_>...Which is a good thing?
16:22<bd_>Peng_: it means the kernel yields the virtual cpu rather than spinning
16:22<JoeK>if i set up htaccess, it will only prompts me for the password once? (after successfully passing credentials)
16:23<Yaakov>No spinlocks are more aware than mine are.
16:23<bd_>Consider: vcpu0 and vcpu1 are assigned to the same physical CPU. vcpu0 takes a spinlock, then is preempted. vcpu1 then spins on the same lock.
16:23<bd_>This is wasteful; both vcpus now need to wait for vcpu1's timeslice to run out
16:24-!-dajhorn [] has quit [Quit: ChatZilla 0.9.85 [Firefox 3.5.4/20091016081620]]
16:24<Peng_>Back in the day, I used Gentoo and compiled my own kernels. Now changing my kernel on Linode scares me. :\
16:24<JshWright>Peng_: so you want everyone else on your host to upgrade as well ;)
16:25<JshWright>speaking of upgrading... back shortly
16:25-!-JshWright [] has quit [Quit: leaving]
16:27-!-ubuntuisloved [] has quit [Ping timeout: 480 seconds]
16:27-!-JshWright [] has joined #linode
16:28<JshWright>well, nothing caught fire yet...
16:28<Peng_>JshWright: You just upgraded?
16:28<JshWright>bd_: I'm a firefighter... _everything_ catches fire eventually
16:28<JshWright>Peng_: yep
16:29<Peng_>JshWright: What's CONFIG_SPONTANEOUS_COMBUSTION set to?
16:29*SelfishMan remembers several laser printers catching on fire
16:29<bd_>JshWright: What, even pure nitrogen gas?
16:29<JshWright>Peng_: I've left it off for now
16:29<linbot>New news from wiki: Ccache <> || Appinfo djbdns <> || AppInfo ccache <> || AppInfo: screen <>
16:29<Peng_>Ehh, I don't want to reboot once a month for kernel upgrades.
16:30<Peng_>OTOH, I need to test that my IPv6 tunnel survives a reboot, and restart NTP.
16:30<SelfishMan>I also remember a chemistry teach at the high school igniting a balloon of hydrogen, then one of oxygen then finally one of acetylene
16:31<Yaakov>tasaro: Sure, break my heart a second time... I lose my beloved Inger and my beloved Newark1 in the same month. I weep, and cannot be consoled.
16:31<JshWright>SelfishMan: mix the O2 balloon with the acetylene balloon...
16:31<jess^>i used to do web design out of a racing garage
16:31<SelfishMan>JshWright: trust me, that wasn't required for the school to get a new chem lab
16:31<jess^>we used to drop baloons full of acetylene off the roof :D
16:32<Peng_>Yaakov: Inger?
16:32<SelfishMan>The guy was smart enough to do all of this under a fume hood but he failed to turn on the ventilation first so there was still plenty of oxygen in there
16:33-!-awnstudio [] has quit [Quit: HydraIRC -> <- It'll be on slashdot one day...]
16:33-!-taylor_ [] has joined #linode
16:34<taylor_>question, does your api allow you to grab bandwith usage?
16:34<linbot>New news from wiki: Vim Tutorial <> || OpenVPN <> || Msec Howto <> || CentOS <>
16:35-!-Everett [] has joined #linode
16:35<SelfishMan>jess^: and what was the point of dropping balloons of acetylene off the roof?
16:35<Peng_>Yaakov: aww.
16:35<jess^>the lit cigarette at the bottom.
16:36<Everett>If no one has a really good answer for this one, I understand, but let me just ask it anyways. I'm used to managed VPS and I have abut 5 sites on managed. A couple of them with heavy traffic (>100,000 pv / day). I've just set up my first linode and I keep just "hanging about" my putty screen and can't seem to take the plunge...
16:37<SelfishMan>Really? You had that good of aim from the roof with a balloon that you could hit a burning cigarette?
16:37<Everett>So, my question is... well basically, is there a final checklist for when to know if your linode is in production shape.
16:37<Everett>I know, it's maybe a silly question but still....
16:37<SelfishMan>Everett: test it.
16:38<Everett>I have, but there's really no easy way of "stress" testing it. at least not that I've found.
16:38<Everett>The thing is, I've moved from Apache to Nginx and moved to PHP on FCGI running php-fpm. Two things I haven't been running before.
16:39<Everett>So, while it's exciting, since the linode is really really fast, it's also a bit scary. What in the world do I do if it breaks :)
16:39<adj>fail back over...
16:39<SelfishMan>ab -n 2000 -c 50
16:39<quanin>fix it?
16:39<SelfishMan>set your ttl to <= 300
16:40-!-darkbeholder [] has joined #linode
16:40<linbot>New news from wiki: Traceroute <> || ZoneEdit <> || Debian Sarge (must be upgraded from Woody) <> || Debian Sarge <> ||
16:40<Everett>yeah, I figure I keep the "old" setup on the managed vps for a little while longer.
16:41<adj>turn down the ttl's, and move the sites ip's one by one
16:41<adj>move them back if you hit a problem
16:41<Everett>adj: yes, that's certainly sounds like a plan.
16:43<Everett>anyone else using nginx here and have good experience with it. It certainly seems stable and veeeery fast, but... well, I have no experience with it from production yet
16:44<SelfishMan>It really isn't the webserver itself that is slow or fast
16:44<SelfishMan>it is the content being served
16:44<SelfishMan>php is just slow
16:44-!-Bohemian [~Bohemian@] has joined #linode
16:45<Everett>SlfishMan: well, true, but there's slow and there's syrup. Moving from my managed setup which was not on XEN has been an eye opener. So relative to what I'm used to it is fast. The webserver nginx is "faster" in terms of not having to load php (mod_php) with every request.
16:45<linbot>New news from wiki: Lish Documentation <> || Linode <> || Kubuntu <>
16:46-!-taylor_ [] has quit [Quit: leaving]
16:47<SelfishMan>Everett: true, but fastcgi can have some scalability liits
16:47<Everett>well, that's sort of what I'm afraid of....
16:50<tylerdu>SelfishMan: care to elaborate (fastcgi scalability limits)
16:51<linbot>New news from wiki: Network <> || Linode Forums <> || Linode DNS <> || How to set up your Linode <> || Edubuntu <
16:51<SelfishMan>tylerdu: depending on the implementation you either have a hard limit for the number of fastcgi processes or you have a high risk of exhausting memory
16:52-!-daMaestro [] has quit [Ping timeout: 480 seconds]
16:52<tylerdu>actually i'm using fastcgi to save the memory (and it works) ;)
16:53<tylerdu>imho, you don't need that many phpcgi processes, because they contend to same CPU resources, so something like 2 x #CPU's is probably enough
16:54<Everett>well, thanks anyways everyone. I'll be switching a site over to the linode in just a while and we'll see how it goes...
16:54<@jed>1 per core, imo
16:54<tylerdu>jed: maybe 2, because if one is stuck on i/o, the other one can use that core
16:54<tylerdu>jed: same thing as make -j x, where x is ncpu + 1 (compiling kernel)
16:54<SelfishMan>it depends on the implementation for memory usage. If multiple processes can share the memory then great but it is pretty common to start a new process for every fcgi instance
16:55<@jed>tylerdu: another can take its place only if PHP yields
16:55<@jed>otherwise it sits in iowait
16:55<tylerdu>SelfishMan: yes, they're separate processes, my point is that there doesn;t need to be that many of them, for instance, on 4 core server, 10 is plenty
16:55-!-Everett [] has quit [Quit: ajax IRC Client]
16:55<@jed>if PHP executes blocking I/O that core is stuck.
16:56-!-daMaestro [] has joined #linode
16:56<tylerdu>jed: the scheduler will yield the process that is stuck on i/o and give cpu to something else that can run
16:56<linbot>New news from wiki: Internal Services <> || Internet Relay Chat Channel <> || Upgrade Linode <> || What Do You Use Linode For? <
16:56<@jed>tylerdu: not always
16:56<tylerdu>it should
16:56<SelfishMan>tylerdu: again, depending on implementation that can limit scalability as you can only have 10 concurrent page views
16:56<Peng_>Bah, Peng_ can't type today.
16:57<SelfishMan>runaway fastcgi processes are pretty common
16:57<tylerdu>SelfishMan: i agree that things get complicated if you need more concurrent accesses, but some kind of buffering probably takes place
16:57<tylerdu>need to investigate that further
16:57<SelfishMan>It comes down to knowing that your scripts are 100% fastcgi safe along with the interpreter
16:58-!-D[a]rkbeholder [] has joined #linode
16:58<tylerdu>fastcgi is just a way of running your .php-s, mod_php is the other one, so it just works after you set up your web server
16:58<SelfishMan>doing *anything* on a production site because you heard or read it was faster is a sure way to break things
16:59<SelfishMan>fastcgi != mod_php
16:59<tylerdu>fastcgi is definitely a way to SAVE on memory, while keeping the performance, i tested that and it really works
16:59<@jed>fastcgi is just an alternate method for the Web server to communicate with PHP, and won't save squat
17:00<@jed>you're observing a red herring
17:00<tylerdu>apache w/ mod_php eats much more memory, because even when your apache is serving favicon.ico, it is served from the process which has php linked in
17:00-!-darkbeholder [] has quit [Ping timeout: 480 seconds]
17:00<SelfishMan>jed: what color is a herring normally?
17:00<spkitty>i figured out how fastcgi saves you memory
17:00<spkitty>it just doesn't do anything
17:00<SelfishMan>!f what color is a herring?
17:00<spkitty>i had to go back to mod_php
17:00<linbot>SelfishMan: This was a triumph. I'm making a note here: HUGE SUCCESS.
17:00<stan_theman>salmon colored?
17:01<tylerdu>it saves memory because your nginx/... server all static content and dispatches only php requests to phpcgi farm
17:01<Guspaz|m>jed: It's not a red herring. If you have a light PHP load and only need one PHP process with fastcgi, rather than with mod_php where you're loading PHP into every Apache process?
17:01<linbot>New news from wiki: Autounmask <> || Apache2 SSL in Ubuntu <> || FAQ <> || Billing Questions <> || BSD Howto <
17:01<Guspaz|m>Although IIRC Apache had some mode to prevent having more than one process.
17:01<@jed>Guspaz|m: loading implies a lack of persistence
17:01<SelfishMan>tylerdu: apache without mod_php is actually quite small
17:01<tylerdu>SelfishMan: i know, but i can't serve php with it
17:01<SelfishMan>also, if you are going to compare memory footprint straight across then use MPM worker instead of prefork
17:01<@jed>apache processes are persistent, subject to configuration knobs, so PHP stays in memory just like with fastcgi
17:01<Guspaz|m>jed: Your point? One persistent copy of PHP with fastcgi versus many persistent copies with mod_php?
17:02<@jed>Guspaz|m: you're bottlenecking yourself running one persistent copy of PHP
17:02-!-hercynium [] has joined #linode
17:02<SelfishMan>Guspaz|m: each fastcgi process typically loads a copy of PHP so there is no gain there
17:02<Guspaz|m>jed: It's an example. You can scale it to your needs.
17:02<@jed>many persistent copies of PHP with fastcgi versus many persistent copies with mod_php?
17:02<Guspaz|m>If I've got Apache with twenty processes serving stuff up, that's 20 copies of PHP. What if I only need frive?
17:02<Guspaz|m>If I needed frive, I'd be fucked.
17:03<@jed>then you're doing a whole lot more static serving than PHP, and I'd be impressed as that's now how PHP operates
17:03<tylerdu>jed: a few phpcgi processes vs few hundred mod_php processes is a HUGE difference
17:03<tylerdu> memory usage
17:03<@jed>tylerdu: a few hundred?
17:03<tylerdu>jed: you have never be slashdotted?
17:03<@jed>you're administering PHP wrong if it gets that high in processes.
17:03<@jed>pardon, Apache
17:03<jess^>i read 'administering' as 'masturbating' out of the corner of my eye
17:03<jess^>and went 'wait, what the f-- oh.'
17:03<@jed>Apache should never, I repeat never, spawn more than 10 processes
17:03<@jed>queue your connections and stay alive
17:03<tylerdu>jed: when you finish on a high profile site, apache forks 250 processes, all with php linked in
17:04<Guspaz|m>Another example, Otakuthon needs both Perl and PHP. But we have far higher PHP load than perl. With Apache's module system, we'd have equal numbers of copies of both. With fastcgi, we can have more PHP processes than perl processes.
17:04*SelfishMan points at keepalives
17:04<tylerdu>jed: fastcgi solution is still 5-10 processes which handle the same load, it has been tested
17:04<@jed>tylerdu: I work at Linode, we have a pretty high-profile site
17:04<@jed>I don't know if you've heard of it :)
17:04<tylerdu>sure i did
17:04<Guspaz|m>jed: You run Cold Fusion :P
17:04<tylerdu>high-profile is many requests per second, do you have that?
17:04<@jed>Guspaz|m: you are aware ColdFusion runs on Apache, correct?
17:05<Guspaz|m>But it's not PHP or Perl.
17:05<path>and phpbb runs on coldfusion?
17:05<tylerdu>many complex pageviews (1 page = 1 php + 10 pics + 5 js + ...)
17:05<SelfishMan>So basically, what you guys are saying is that nginx proxying to apache with mod_php will have the same memory footprint as nginx proxying to fastcgi
17:05<path>and the mostly defunct wiki
17:05<Guspaz|m>jed: They're not terribly high traffic.
17:05<tylerdu>SelfishMan: that is correct
17:05<@jed>Guspaz|m: but they're PHP
17:06<@jed>and run on the same Apache
17:06<Peng_>The wiki too.
17:06<path>and the pastebin
17:06<Guspaz|m>My point is, fastcgi lets you tune the amount of memory spent on a given resource (static, PHP, perl, etc) based on demand, while Apache just loads everything into every process.
17:06<linbot>New news from wiki: Ubuntu/installing Apache and PHP5 with MySQL support <> || Ubuntu Server - Using NMAP to scan your server for vulnerabilities. <> || Elogv <
17:06<SelfishMan>tylerdu: so basically, you're saying that fastcgi is no different than apache in memory usage
17:06<Peng_>Oh, path said the wiki already. Oops.
17:06<path>so did linbot
17:06<tylerdu>SelfishMan: it is, in your setup you're proxy ONLY php requests to apache behind, so it forks much less processes
17:06*SelfishMan glares at pparadis_
17:06<Guspaz|m>We only have our reg system running on Perl while we have everything else on PHP. Clearly we need more PHP than perl. Why load Perl into every process?
17:06<@jed>SelfishMan: stan_theman
17:07<SelfishMan>jed: THE INTERN?
17:07<tylerdu>only Guspaz|m here understands what i understand (and have tested in practice) :(
17:07<SelfishMan>You game him access to the wiki?!?
17:07<spkitty>while we're speaking of apache - does anybody know why it will occasionally take *minutes* to connect to my site, but othertimes instantaneously
17:07*stan_theman laughs maniacally
17:07<Guspaz|m>jed: Even if we had 10 processes all told, I'd have 10 copies of Apache, 10 copies of PHP, and 10 copies of Perl. When in reality, assume that I have a fastcgi setup, I might have 10 copies of Apache, 10 copies of PHP, and 3 copies of perl.
17:07<SelfishMan>spkitty: *hint*keepalives*hint*
17:08<Guspaz|m>Although I use lighttpd, so I only have the one process for static content.
17:08<Guspaz|m>I fail to see how loading way more copies of perl than we need could possibly use the same amount of RAM.
17:08<@jed>Guspaz|m: what's your advantage?
17:08<SelfishMan>unless you are running a few very screwed up php modules you can run php with no problems under mpm worker
17:08<SelfishMan>which makes all this moot
17:09<@jed>SpaceHobo: shh, that's my hole card
17:09<spkitty>what should i be changing requests and timeout too then SelfishMan
17:09<spkitty>say for a base model linode
17:09<Guspaz|m>jed: Errm, my advantage is that we save memory by not having more copies of Perl that we don't need?
17:09<@jed>define "copy of perl"
17:09<SelfishMan>spkitty: KeepAlives Off
17:09*SelfishMan hands jed the camel book
17:09-!-linville [] has quit [Quit: Leaving]
17:09<Guspaz|m>jed: Does mod_perl not behave the same as mod_php and load a copy of perl and various libraries into every Apache process?
17:09-!-Bohemian [~Bohemian@] has quit [Quit: Bohemian]
17:10<tylerdu>mod_perl is especially hungry of memory
17:11<tylerdu>but also very powerful
17:11<Guspaz|m>perl is hungry for memory in general, which is something of a problem for us, hence the need to keep the number of instances of perl under control. Our web guy is using some horrible framework that for some reason loads in practically every possible perl module.
17:11<linbot>New news from wiki: NetworkManagerService <> || RTFM <> || Rebooting on OOM <> || VML <> || PV-GRUB <
17:11<SelfishMan>tylerdu: look at the memory footprint of apache prefork with just mod_php and compare that to just mod_perl and see how your statement stands
17:12<tylerdu>i have both, and mod_perl equipped apache processes are MUCH bigger
17:12-!-orudie [] has quit [Quit: Leaving]
17:13<SelfishMan>Really? cuz I have an instance on one node using only mod_perl that uses 1.8MB and an instance on another node with only mod_php that uses 14.4MB
17:13<Guspaz|m>My perl processes use up ~50MB of RAM a pop, while my php processes use ~20MB a pop. All damned to that bloody framework the guy uses.
17:14<tylerdu>SelfishMan: that is odd, i think you can't have such a small mod_perl process
17:14<SelfishMan>Guspaz|m: that isn't really the fault of perl
17:14<Guspaz|m>Although it appears that somebody has reduced the number of PHP children to 2, and that worries me tremendously.
17:14<linbot>SelfishMan: Yo momma's so ignorant, She thinks a computer is the monitor. (758:0/2) [mrmou]
17:14<SelfishMan>well, !urmom is running under mod_perl with only 1.8MB in use
17:14<tylerdu>Guspaz|m: those 2 PHP children still can serve many requests
17:14<SelfishMan>seems to be working for me
17:15<spkitty>SelfishMan: i have maxclients set to 150 and there's no way i'm hitting that limit (probably a good thing too since i just realised that would put me heavy into swap), you really reckon i should turn keepalive off?
17:15<tylerdu>well, my local apache has ~ 20MB RSS, that is with both mod_php and mod_perl loaded
17:15<@jed>that's an absurdly high maxclients
17:15<Guspaz|m>tylerdu: And when phpbb2 starts racking up the simultaneous sessions?
17:15<Guspaz|m>Or whichever forum we use.
17:16<tylerdu>Guspaz|m: should work, it is technically cpu bound
17:16<tylerdu>when you run out of cpu power it will fail
17:16<SelfishMan>spkitty: turn the maxclients down too but yes, turn off kepalives
17:16<linbot>New news from wiki: Moving from Debian to Ubuntu <> || Custom Distribution <> || Conventions used in this tutorial <> || Conf-update <
17:16<spkitty>okay thanks
17:16<Guspaz|m>And if we're DB-bound, and two PHP processes are tied up waiting for a DB query when another request comes in for something that doesn't touch the DB?
17:17<tylerdu>Guspaz|m: in that case i'd suggest to increase number of processes, although i think db is much faster than php processing
17:17<tylerdu>especially if you use the real database like postgresql
17:17<Guspaz|m>We're using less than half the RAM on a 540, and our budget could support a larger linode if required, I've no idea why he decreased it from 6 to 2.
17:18<Guspaz|m>(he being the other admin)
17:18<tylerdu>typically your db, if optimized, can handle hundreds or thousands queries per second, but your favorite cms framework will be able to push only few php pages per second
17:18<tylerdu>Guspaz|m: on linode with 4 cpus, i'd suggest like 8 phpcgi processes
17:18<Yaakov>My favorite CMS network is C.
17:18<Yaakov>err.. framework.
17:18<tylerdu>each one costs about 10MB od memory, iirc, and that way you can be sure to use all available cpu cores if needed
17:19<tylerdu>Yaakov: good choice :)
17:19<tylerdu>Yaakov: do you, by any chance, have drupal rewritten in C? ;)
17:19<Guspaz|m>Well, we should have at least 4 children, not 2, then.
17:19<Yaakov>Actually it looks like we are going to stand up Plone internally.
17:19<Guspaz|m>Although I've never been clear, does the parent PHP process actually serve requests?
17:19<tylerdu>Guspaz|m: yes, technically at least 4
17:20<Guspaz|m>As in, I've got one process with two children, so three processes. But the parent process never seems to use nearly as much RAM as the children, so I assume it doesn't actually handle requests?
17:20<tylerdu>but as i said, if one of them on one core gets stuck on i/o (very probable) the other one will at least have the chance to utilize that core
17:21<tylerdu>although jed thinks that is not guaranteed to work that way, i think it is, when process is stuck on i/o it is put on a wait queue, and that is an rescheduling point, because there's no use to wait for i/o which is many orders of magnitude slower than cpu
17:22<linbot>New news from wiki: Fstab <> || Useful Ubuntu packages <> || Host SSH Key Fingerprints <> || Getting started <> || Grc <http
17:22<tylerdu>Guspaz|m: i don't know internals, but it looks there's a dispatcher and workers, so if you start 4 php-cgi workers, you'll in fact have 5 processes, but you'll see that only 4 of them burn cpu cycles
17:22<Guspaz|m>Hrrm linbot is geting rather spammy with wiki updates.
17:22<linbot>SelfishMan: Spam subject of the hour: you need to check your Bank Deposit Insurance Coverage
17:23<tylerdu>!xkcd penis
17:23<linbot>tylerdu: The Sierpinski Penis Game:
17:23-!-compwhizii [] has quit [Read error: Connection reset by peer]
17:24<tylerdu>geocities is no more :)
17:24-!-Bohemian [] has joined #linode
17:25<SelfishMan>I haven't been to in a really long time.
17:25*SelfishMan pets RSS
17:25<tylerdu>what rss reader do you use?
17:25<Guspaz|m>X10KCD SPY CAMERA!!!! (TESTED ON KITES)
17:26-!-Bohemian [] has left #linode []
17:26<Guspaz|m>Wow, never thought I'd see "<SCRIPT LANGUAGE='SCHEME'>"
17:27-!-bogdanbiv [~quassel@] has quit [Ping timeout: 480 seconds]
17:27-!-huseyin [] has quit [Quit: Reconnecting]
17:27<linbot>New news from wiki: Bandwidth <> || Debian libc6 and 2.6.x kernels <> || Finnix LiveCD Recovery Distribution <> || IO Tokens <
17:29<tarpman>Guspaz|m: does that actually work in any browsers?
17:30<Peng_>SelfishMan: !spam is so going to get linbot akilled someday.
17:31<Guspaz|m>I don't think so?
17:32<linbot>New news from wiki: Category:Linux Applications <> || Debian Sarge (must be upgraded from Woody) <> || Debian Sarge <> || Category:Stub <
17:32<tylerdu>ha ha ha
17:33<Guspaz|m>caker: Is it possible to turn off wiki updates on linbot?
17:33<tylerdu>linbot: turn off wiki updates
17:33<@caker>of course it's possible, yes
17:34<jess^>my god
17:34<jess^>this customer is... argh. annoying me.
17:34<linbot>I'm sorry, tylerdu, but I'm afraid I can't do that.
17:34<tylerdu>linbot: j/k ;)
17:39<linbot>New news from wiki: Multiple IPs <> || Category:Troubleshooting <> || Category:Linode Specific <> || Category:Distros <
17:39<tarpman>stan_theman: busy busy
17:40<stan_theman>Most of the pain should be over now
17:40<Peng_>Oh no. I like doing a backup before a kernel update, and just like last time, my crappy backup server is being horribly slow. :(
17:41<Yaakov>caker: Is it possible to fit 25 pounds of monkeys in to a 5-pound monkey bucket?
17:41-!-Sputnik7 [~Sputnik7@] has quit [Quit: -=SysReset 2.53=-]
17:42<linbot>New specials on dinner menu: 12oz prime rib for $9.95, with soup or salad, and your choice of starch and vegetable || braised salmon in a merlot reduction for $13.95, also with soup or salad and your choice of starch and vegetable || eggplant parmesan served over linguini for $11.95, with your choice of soup or salad
17:42<Peng_>What grade of beef is the prime rib?
17:42<Peng_>I've probably asked this before, but: if I want to shut down my 'node without it being rebooted, what should I do?
17:43<stan_theman>Peng_: Grade A, but the caveat being that I can edit it at anytime :)
17:43<Peng_>Well, I can press the Shutdown button in the dashboard, but I also want to twiddle a couple things inside my node before doing it.
17:43<Peng_>Yaakov: Lassie won't kick in?
17:43<Yaakov>I don't know!
17:43<HoopyCat>Peng_: you can do it via a linode.shutdown API call, or by ssh shutdown
17:44<quanin>or just ssh into lish, do what you gotta do on the node, detach and shutdown.
17:45<HoopyCat>Yaakov: i left my sound card in san francisco
17:45<Yaakov>HoopyCat: You stink on ice.
17:45*HoopyCat beats Yaakov around a bit with a 74HCT08
17:46<SelfishMan>Peng_: quite possibly
17:46<Peng_>...What's the difference between "halt" and "poweroff"?
17:47*Peng_ Googles!
17:47<HoopyCat>Peng_: in theory, halt will merely halt the system; poweroff will turn the power off before doing so.
17:48<Peng_>OK. What does that mean?
17:49<SelfishMan>HoopyCat: need some 74HCT08s? I have a few rails of them
17:50<bd_>Peng_: halt means the kernel goes into an infinite loop. poweroff means it goes and turns off the PSU :)
17:50<Guspaz|m>A sound card? People still have those?
17:50<HoopyCat>Peng_: on linode, there is no difference. on PC hardware running teh lunix, i believe there is no difference. on other iron, there may be a difference.
17:50<Guspaz|m>Soundcards became mostly pointless when everything moved to software. Only difference now is the SNR, and a lot of onboard stuff is pretty decent.
17:52<HoopyCat>SelfishMan: naw, i have enough for the current project. just got the mouser shipment in
17:53<Peng_>bd_: Halting sounds silly, then?
17:53<Peng_>HoopyCat: OK.
17:54<HoopyCat>Peng_: not if your server is the size of a small village and requires a plethora of minutes to get everything spun up, and you just want to halt the kernel for a bit so you can warm-swap a detachable jesus or something :-)
17:55<linbot>New news from wiki: Swappiness <> || Traceroute <>
17:55<Peng_>HoopyCat: Oh, okay.
17:56<Peng_>Crap! I left rsync on "Enter passphrase for key" until it died! :(
17:57<HoopyCat>warning: do not look directly into "Kingbright LED Super Bright" with remaining eye
17:59-!-hpj [] has quit [Read error: No route to host]
18:00<bd_>Peng_: Well, you have to remember, people didn't always have these fancy-shmancy 'ACPI' thingamabobbers
18:00-!-awnstudio [] has joined #linode
18:01<bd_>back then there was only 'halt' and 'reboot'
18:01<bd_>and 'reboot' was considered a luxury!
18:01<metap|pe>it is now safe to turn off your computer
18:01<bd_>You kids these days have it easy.
18:01<HoopyCat>in practice, poweroff sent "Please turn the FOOVAX1 server off at your next convenience. Thanks!" to the line printer before halting
18:01<Guspaz|m>Good old AT.
18:05<tarpman>Yaakov: until recently a guy worked with me who always did three syncs before a shutdown or reboot. I never could figure out why.
18:06<Yaakov>tarpman: He probably started on Solaris.
18:06<Yaakov>Or, BDSI.
18:06<Yaakov>Or maybe AIX or HP-UX.
18:07<Yaakov>I wonder if I can get SCO to run on a 'node.
18:07-!-ericoc [] has joined #linode
18:07<straterra>I doubt it
18:08<Yaakov>You drive a Datsun, what would you nkow?
18:08-!-det [] has quit [Remote host closed the connection]
18:08<straterra>I don't drive a Datsun
18:08<straterra>You insensitive clod
18:09<Eman>what about a chrysler <lisp><wrist type=flick>lebaron</wrist></lisp>
18:10<Yaakov>straterra drives a Datsun B210, I think.
18:10<Guspaz|m>I PREFER THE Z
18:12<straterra>i used to own a chrysler
18:14-!-ubuntuisloved [] has joined #linode
18:15-!-Guspaz|m [] has quit [Quit: ajax IRC Client]
18:18<Peng_>Oh my god. I just did it _again_.
18:19<Peng_>Oh, right. It's because I always forget I have to enter the password _twice_: Once for sudo and once for ssh.
18:19<straterra>Use an SSH key..and set up sudo properly
18:20<Peng_>straterra: I do use SSH keys. sudo doesn't get along with ssh-agent and I've never fixed it.
18:20<Peng_>straterra: What's wrong with my sudo setup?
18:20<straterra>You have to enter a password
18:21<Peng_>What, you like passwordless sudo?
18:21<HoopyCat>don't turn a backup process... into a backed-up process! set it, and forget it!
18:24<straterra>I do know I fucking hate security torx
18:24<HoopyCat>my backup server has a passphraseless key that is allowed to log in to the client machines for a specific command only; that command is sudo, which allows a passwordless sudo for the specific rsync command line it needs. this does mean i have to uncomment a couple lines to allow a restore to work, but that's not necessarily a bad thing
18:26<kenichi>i learned to drive stick on a B210 with a "super pretzel loves you" bumper sticker.
18:26<kenichi>just fyi
18:28<Yaakov>straterra: Security Torx is nothing...
18:28-!-Redgore [] has quit [Quit: Peace and Protection 4.22.2]
18:28<straterra>I hate those
18:28<straterra>I don't need those
18:28<straterra>Our server room LOCKS
18:30<Yaakov>Google uses them on their appliances...
18:30<Yaakov>So when you want to reverse engineer one you have to get them out.
18:30<straterra>They need to die
18:30<straterra>Dont they sell the bits to anyone?
18:30<Yaakov>In 10 seconds, with no wrench.
18:30<straterra>So..WTF is the point?
18:31<Yaakov>They thought it would stop me.
18:31<Yaakov>Probably would have stopped other people.
18:31<stan_theman>"Google Snatchers", starring Yaakov!
18:31-!-FooMunki [] has joined #linode
18:32<Yaakov>I have defeated many varieties of fasteners.
18:33<jess^>them: "if i connect to my server from the outside and send a test message from my yahoo account to my account at $DOMAIN_WHICH_THE_BOX_HOSTS_MAIL_FOR, it gets accepted without prompting me for a username and password!"
18:33<jess^>them, continued: "can you please fix it so that it requires a username and password for people to send mail?"
18:33<jess^>me: *snrf*
18:39<linbot>New news from wiki: Main Page <>
18:45<linbot>New news from wiki: Main Page <>
18:51<linbot>New news from wiki: Main Page <>
18:51<straterra>jess^: wtf
18:53-!-chemosh [] has joined #linode
18:55*mikegrb puts Yaakov on the do not sell linode appliance to list
18:58<Internat>whats the easiest way in bash to rename all *.txt files to *.TXT
18:58<straterra>with a for statement
18:59<BarkerJr>really? I would think of using the find command
19:00<bd_>debian has a rename utility :)
19:00<Internat>sorry rename all within one folder
19:00<bd_>rename 's/\.txt$/.TXT/' *.txt
19:00<Internat>ls | xargs something?
19:00<jess^>straterra: the guy is upset because he can telnet to his server (say and send an email from anywhere to himself at and the mailserver accepts it with no password
19:00<bd_>note: other distributions have rename utilities that do very different things
19:00<straterra>jess^: uhm..yes
19:00<straterra>Thats how email works
19:01<jess^>straterra: exactly
19:01<Yaakov>Email works?!
19:01<Yaakov>Who knew?
19:01<jess^>straterra: i'm laughing because this guy "has been in the computer industry for 20 years and has never seen an e-mail system so insecure!"
19:02<jess^>i seriously was laughing so hard i had to run to the bathroom so i wouldn't piss myself laughing
19:02<BarkerJr>who is Sschwertly? :)
19:02<Yaakov>jess^: I am not following thing.
19:02<Yaakov>jess^: I am not following this.
19:02<Internat>pretty sure thats how email works lol
19:02<Internat>Yaakov, the guys complaining cause he can telnet to his mail server, and send himself an email from any address
19:03<Yaakov>He is spoofing the conversation?
19:03<jess^>and that the mailsystem accepts it without issue.
19:03<jess^>no, just using a from address that he has elsewhere
19:03<Internat>Yaakov.. no, i imagine hes sending a proper converstation
19:03<Yaakov>telnet 25 ?
19:03<straterra>He wants people to have to put in a username/password to email to him?
19:03<jess^>apparently so
19:04<Internat>Yaakov. because its the recieving server for it HAS to accept any email that connects to it for him.
19:04<Yaakov>Set up SPF and DKIM.
19:04<Internat>well, it could do spam, and sender/blacklist filtering, but thats beside the point
19:04<Internat>Yaakov: no. that would be the senders servers thing, not the receiever
19:04<straterra>spf is the reciever
19:04<Yaakov>No, it would mean that he couldn't send mail from as he can now.
19:05<Yaakov>And DKIM would confirm the headers in the email.
19:05<Internat>that assumes that the domain he is sending from, has spf records, and dkim
19:05<Yaakov>yahoo does.
19:05<Internat>but we never said yahoo.
19:05<Yaakov>< jess^> them: "if i connect to my server from the outside and send a test
19:05<Yaakov> message from my yahoo account to my account at $DOMAIN_WHICH_THE_BOX_HOSTS_MAIL_FOR, it
19:06<Yaakov> gets accepted without prompting me for a username and password!"
19:06<Internat>oh, i missed the part about yahoo, it had scrolled off my screen.
19:06-!-Desph [] has quit [Quit: Desph]
19:07<Internat>08:59] <jess^> straterra: the guy is upset because he can telnet to his server (say and send an email from anywhere to himself at and the mailserver accepts it with no password
19:07<linbot>New news from wiki: Support ticket <> || Support Ticket <>
19:07<Yaakov> < Yaakov> No, it would mean that he couldn't send mail from as he can now.
19:07<Yaakov>(I just wanted to do that too.)
19:08<Internat>yeah, that would mean that he couldnt send from yahoo.. it would fail.. but the rest of the domains that dont have spf or dkim would still pass fine
19:08<Yaakov>But he would be happy.
19:08<Yaakov>Unless you sowed discontent by having him check a non-spf domain.
19:08<jess^>i LOLed so hard
19:09<linbot>Yaakov: [spf] none: does not designate permitted sender hosts (spf record: )
19:10<Yaakov>Heh, yahoo doesn't use spf.
19:10<Yaakov>It seems.
19:10<Yaakov>But... they demand it, I think.
19:10<Yaakov>Doesn't seem right.
19:10<Yaakov>!dig TXT
19:10<linbot>Yaakov: [dig] status: NOERROR | ;; ANSWER SECTION: | ;; AUTHORITY SECTION 521 IN SOA 2009102613 3600 300 1814400 600
19:10<Yaakov>!dig MX
19:11<linbot>Yaakov: [dig] status: NOERROR | ;; ANSWER SECTION: 7053 IN MX 1 | ;; AUTHORITY SECTION 163192 IN NS 163192 IN NS 163192 IN NS 163192 IN NS 163192 IN NS 163192 IN NS 163192 IN NS
19:11<Yaakov>Oh well.
19:13<chemosh>Hi. I have a directory thats 183MB. When rsync-ing it to my linode its 198MB. Without any changes to the source, increment dirs with --link-dest become 15MB. Any idea where the extra 15MB comes from? Locally it works fine and the increments dirs are only a few KB.
19:14<Jonathan1>chemosh: Just a wild guess, but block size of the fs differs?
19:14-!-jonny5 [] has joined #linode
19:14<jonny5>Welcome jonny5
19:15<Yaakov>A little solipsistic...
19:15<Peng_>jonny5 refers to himself in the third person?
19:15<Jonathan1>or it was a command
19:16<jonny5>You will welcome me!
19:17<jess^>KNEEL BEFORE ZOD
19:18<jonny5>Kernel before Lord Zod
19:20<Thorgrimr>Oh, okay... Is it installable via apt on there?
19:21-!-Bohemian [] has joined #linode
19:21-!-Bohemian [] has left #linode []
19:21<chemosh>Johanthan1: its 4k @ linode, HFS+ should be aswell
19:22<Jonathan1>chemosh: My only other wild guess would involve mythical beings, no idea
19:24<chemosh>k. Cheers :)
19:24-!-kenichi [~kenichi@] has quit [Quit: kenichi]
19:26<jess^>bun burger lettuce onion tomato salt pepper + + + + + + eat
19:27-!-kelvinq [] has joined #linode
19:30<jonny5>ketchup mustard
19:32<jonny5>Anyone got Windows 7 yet?
19:32<bob2>lots o people
19:32<bob2>since the rc was freely available for months
19:33-!-darkbeholder [] has joined #linode
19:37-!-lakin [] has joined #linode
19:37-!-D[a]rkbeholder [] has quit [Ping timeout: 480 seconds]
19:41-!-memeverse [] has joined #linode
19:47<linbot>New news from wiki: Main Page <>
19:48-!-edonemem [] has quit [Ping timeout: 480 seconds]
19:50<HoopyCat>*puts the spanner down* alrighty, does properly redirect properly and load properly?
19:52<HoopyCat>hooray, dynamic munin graph generation without breaking existing links. got a regexp right on the first try, nice.
19:53<Peng_>I want to see it 12% full. :P
19:53<linbot>New news from wiki: Main Page <>
19:54<HoopyCat>Peng_: if it goes over the top capacitor, it's a short, so the charging timer locks up
19:54<Peng_>That's no fun.
19:55<Peng_>HoopyCat: Do you have nagios running in case that happens? :D
19:55<HoopyCat>Peng_: no, i just notice that the LED that lights during the discharge cycle stops flashing :-)
19:55-!-jonny5 [] has quit [Remote host closed the connection]
19:56-!-FooMunki_ [] has joined #linode
20:01<Pryon>HoopyCat: do you have any way of muninizing local relatie humidity?
20:01-!-FooMunki [] has quit [Ping timeout: 480 seconds]
20:01-!-FooMunki_ is now known as FooMunki
20:02<Pryon>I see the purchase of a weather station in your immediate future
20:02<HoopyCat>Pryon: with the right module, sure
20:07-!-hfb [] has quit [Quit: Leaving]
20:07-!-daMaestro [] has quit [Remote host closed the connection]
20:08<BarkerJr>there's a 360 on fremont!
20:08<BarkerJr>for those who have been waiting
20:17-!-Turl [] has joined #linode
20:18-!-bryanl [] has joined #linode
20:19-!-lakin [] has quit [Ping timeout: 480 seconds]
20:25<SelfishMan>BarkerJr: old news. fremont doesn't have shortage problems all that much anymore
20:25<linbot>Peng_: Fremont360 - 1, Fremont540 - 26, Fremont720 - 8, Fremont1080 - 6, Fremont1440 - 5, Fremont2880 - 1, Fremont5760 - 1, Fremont8640 - 1, Fremont11520 - 1, Fremont14400 - 1
20:25<Peng_>It's rather short now.
20:25<linbot>Peng_: Dallas360 - 59, Dallas540 - 35, Dallas720 - 22, Dallas1080 - 18, Dallas1440 - 13, Dallas2880 - 2, Dallas5760 - 1, Dallas8640 - 1, Dallas11520 - 1, Dallas14400 - 1
20:25<linbot>Peng_: Newark360 - 22, Newark540 - 28, Newark720 - 24, Newark1080 - 7, Newark1440 - 5, Newark2880 - 3, Newark5760 - 2, Newark8640 - 1, Newark11520 - 1, Newark14400 - 1
20:25<linbot>Peng_: Atlanta360 - 38, Atlanta540 - 26, Atlanta720 - 19, Atlanta1080 - 23, Atlanta1440 - 17, Atlanta2880 - 2, Atlanta5760 - 1, Atlanta8640 - 1, Atlanta11520 - 1, Atlanta14400 - 1
20:26<Peng_>Sorry. :P
20:26<linbot>bd_: Availability (all): Atlanta360-38 Atlanta540-26 Atlanta720-19 Atlanta1080-23 Atlanta1440-17 Atlanta2880-1; Dallas360-59 Dallas540-35 Dallas720-22 Dallas1080-18 Dallas1440-14 Dallas2880-1; Fremont360-1 Fremont540-26 Fremont720-8 Fremont1080-6 Fremont1440-5 Fremont2880-1; Newark360-22 Newark540-28 Newark720-24 Newark1080-7 Newark1440-5 Newark2880-2; (1.02314) urmom says hi
20:26*SelfishMan slaps the fsck ou of Peng_
20:26<linbot>SelfishMan: Spam subject of the hour: you need to check your Bank Deposit Insurance Coverage
20:27<linbot>bd_: (ham <US zip code | US/Canada city, state | Foreign city, country>) -- Returns the approximate weather conditions for a given city.
20:27<bd_>!ham 01003
20:27-!-nenolod [] has joined #linode
20:27<chemosh>Johathan1: If I tarball the remote dir and compare it locally, its the same size. Seems you're right. Weird.
20:27<bd_>No such location?
20:27<bd_>I AM LOST
20:28<Trystan>poor thing
20:28<SelfishMan>bd_: YOU FAIL
20:28<SelfishMan>!f weather absecon new jersey
20:28<linbot>SelfishMan: temperature | 50 deg F (wind chill: 48 deg F); conditions | clear; relative humidity | 87% (dew point: 46 deg F); wind speed | 6 mph; (35 minutes ago)
20:28<bd_>!newercalc 01003
20:28<linbot>bd_: 1 thousand and 3
20:28<bd_>!newercalc 01003 zip code
20:28<linbot>bd_: location | Amherst,Massachusetts,United States; county | Hampshire County, Massachusetts
20:28-!-lakin [] has joined #linode
20:28<SelfishMan>!f weather 01003
20:28<linbot>SelfishMan: All your base are belong to us
20:29<linbot> <-- push it. now.
20:33<Peng_>Can SSH tell me which line in known_hosts corresponds to a specific server?
20:34<Peng_>It tells me when it thinks the server is a fraud.
20:34<ericoc>it says what line number in the error usually?
20:34<Peng_>ericoc: Yes. This is not an error condition. I just want to know the line.
20:34<Peng_>(I want to copy it to another machine.)
20:35<bd_>Peng_: There's an option to disable host masking - but it won't affect lines entered prior to enabling it
20:35<ericoc>i'd love to know too tbh
20:35<Peng_>Is it just me, or does OpenSSH go out of its way to make it difficult to verify hosts?
20:36<Peng_>I'm asking cuz I want to copy one line of known_hosts from one machine to another, but I don't know which line it is and don't really want to try them one at a time.
20:36<bd_>ssh-keygen -l -f /etc/ssh/
20:36<bd_>at ssh -v, it'll show something like:
20:36<bd_>debug1: Host '' is known and matches the RSA host key.
20:36<bd_>debug1: Found key in /home/bd/.ssh/known_hosts:2
20:36<Peng_>Ooh, good suggestion.
20:37-!-chemosh [] has quit [Remote host closed the connection]
20:37<Peng_>bd_: Thanks a lot. The ssh-keygen thing was something I had been wanting to know too.
20:37<bd_>Also, disable HashKnownHosts in your .ssh/config to have them in plaintext from now on
20:37<bd_>note that the default on is set by debian, not by openssh :)
20:38<ericoc>default on Arch as well i guess
20:38<Peng_>bd_: <333
20:38<bd_>the idea, I believe, is to slow down worms that use /home/*/.ssh/known_hosts to find targets for ssh bruteforcing
20:38<Peng_>I'll leave HashKnownHosts on, though. It doesn't bother me much.
20:39<Peng_>As long as I can work around it when I need to.
20:40-!-laser` [] has quit [Remote host closed the connection]
20:54-!-kassah [] has quit [Quit: Leaving]
20:55-!-kassah [] has joined #linode
20:56-!-kassah [] has quit []
20:56-!-Berto [] has joined #linode
20:58-!-kassah [] has joined #linode
21:03-!-sc0field [~rajiv@] has joined #linode
21:06-!-blognewb [~blognewb@] has joined #linode
21:07-!-Turl [] has quit [Quit: Leaving.]
21:14-!-kassah [] has quit [Ping timeout: 480 seconds]
21:16-!-BarkerJr [~BarkerJr@] has quit [Remote host closed the connection]
21:18-!-atula [] has joined #linode
21:18-!-OverlordQ [] has joined #linode
21:25-!-kelvinq [] has quit [Quit: kelvinq]
21:28-!-nenolod [] has quit [Quit: Leaving]
21:30-!-kassah [] has joined #linode
21:35-!-kassah [] has quit [Read error: Connection reset by peer]
21:38*caker waves goodbye to GeoCities
21:38<@caker> :)
21:42-!-elhippo [] has quit [Ping timeout: 480 seconds]
21:46-!-pharaun [] has quit [Read error: Connection reset by peer]
21:46-!-lesterc [] has joined #linode
21:47-!-memeverse [] has quit [Remote host closed the connection]
21:49<MJCS> bye bye
21:50-!-memenode [] has joined #linode
21:53-!-Bohemian [] has joined #linode
21:53-!-pharaun [] has joined #linode
21:54-!-Bohemian [] has left #linode []
21:55-!-FooMunki [] has quit [Quit: FooMunki]
21:57<OverlordQ>but I agree with the first comment "I wonder how many people actually got that"
21:59<X-LP>they dont say what is offered with their $5/mo webhosting
21:59<X-LP>ok yeah they do
21:59<KingTarquin>I keep getting loads of "internal dummy connection" messages in my apache log
21:59<@jed>KingTarquin: normal
22:00<KingTarquin>I never used to get them on my other webserver (cant remember the version of apache it used)
22:07<X-LP>geocities tribute!
22:10<@caker><HTML WEB="2.0">
22:11<X-LP>at the bottom it says its best viewed with netscape navigator
22:11<ajmitch>caker: you missed the lisp
22:11<ericoc>he should have linked the w3 validator button
22:11<@caker>aahhh, scheme
22:11<metap|pe>navigator gold, i hope
22:11<straterra>navigator gold was awesome
22:12<metap|pe>it was gold. and not navigator 4
22:13<@jed>netscape was the best browser ver
22:13<@jed>true story
22:14<@jed>navigator...4.61 or so, was the golden age
22:14<ericoc>Best viewed in Netscape Navigator 4.7 or better at 800x480
22:15<ajmitch>must be for a mobile device :)
22:18*X-LP likes the broken link to pictures icon
22:19-!-Keith-BlindUser [] has joined #linode
22:19<metap|pe>for those that could not do the ftp
22:19<Keith-BlindUser>So the new Linode kernels don't show a particular percent of the CPU when idle? Why is that.
22:21-!-sketchyd [] has joined #linode
22:21<sketchyd>hi all
22:21<MJCS>rofl on the numa numa
22:22-!-tomg [] has joined #linode
22:23-!-sketchyd_ [] has joined #linode
22:24<tomg>hi. how do you setup multicast?
22:24<sketchyd_>hi, hopefully easy question - i set up a few virtual hosts, and i want to verify that they are working correctly before i move the domains over
22:24<sketchyd_>for other sites, ive been able to hit the ip directly, but that doesnt seem to be working here
22:25<sketchyd_>what am i missing
22:25<sketchyd_>when i hit my ip, i get a message that says "it works" and thats it
22:25<Trystan>well the IP isnt a virtual domain
22:25<jess^>you need a third level domain you can test with
22:25<jess^>or something like that :D
22:26<metap|pe>"it works" is the default page
22:26<Trystan>sketchyd: you could set your hosts file to say that the domain exists on that IP
22:26<Trystan>thus, skipping the DNS step and letting you test
22:26<jess^>or that
22:26<sketchyd_>trystan - that sounds nice
22:26-!-sketchyd [] has quit [Remote host closed the connection]
22:27<metap|pe>we could have told him to renuberate his glue records and gotten the same result
22:27<sketchyd_>i followed the setup instructions on the library, but i didnt like the step about configuring virtual hosts...
22:28<metap|pe>what didn't you like about it?
22:29<sketchyd_>it doesnt explain anything, it just tells you what to type
22:29<tomg>need to do multicast on ubuntu 8.04. is there a howto forum post or article? any direction = : )
22:31<sketchyd_>trystan - how do i set that up?
22:31-!-tsp [] has quit []
22:32<Trystan>sketchyd_: what operating system are you using?
22:33<Trystan>on your local machine?
22:33<sketchyd_>on my local i have windows vista
22:33<Trystan>and the easier method, before I suggest this one
22:33<sketchyd_>im all set up w/ ssh, sftp, mysql, etc
22:33<Trystan>is the server running DNS that points that domain (just not yet in use)
22:33-!-cweilem [] has joined #linode
22:33-!-argoe [~argoe@] has joined #linode
22:33<Trystan>cos if so you could just change your DNS settings to be the server
22:34<Trystan>if not,
22:34<sketchyd_>what do you mean?
22:35<Trystan>in system32\drivers\etc\ there should be a 'hosts' file
22:35<Trystan>open it with notepad
22:35<Trystan>and add a line
22:35<metap|pe>notepad %systemroot%\system32\drivers\etc\hosts
22:35<sketchyd_>id rather not edit my local
22:35<sketchyd_>that seems busted
22:35-!-argoe [~argoe@] has quit []
22:35<Trystan>it is busted, its purposfully busting it just so you can test
22:35-!-argoe [~argoe@] has joined #linode
22:36-!-tsp [] has joined #linode
22:36<sketchyd_>wouldnt it be better to have the site properly hosted on the server before i flip the domain?
22:36<Trystan>this isnt flipping the domain
22:36<Trystan>other people will access it as normal
22:36<sketchyd_>i know its not
22:36<sketchyd_>i want to set up my new server as if it was on a domain
22:36-!-Keith-BlindUser [] has quit [Quit: Freedom Chat - Your Home Away From Home | | tIRC script by the Freedom Chat Leets]
22:36<metap|pe>you're just faking out your machine so your browser sends the virtual domain info to the server
22:36<sketchyd_>yeah, i get that
22:36<sketchyd_>i dont want to do that
22:37<metap|pe>when done, remove entries from hosts
22:37<Trystan>i'm not sure what you are after then
22:37<metap|pe>flip dns
22:37<sketchyd_>theres no way i can hit <my ip>/<my website>
22:37<sketchyd_>and then later, point the domain name at that folder
22:37<Trystan>not to test a virtual host
22:37<sketchyd_>ive done it with every other webhost ive had
22:37<metap|pe>that would not test the virtual host
22:38<sketchyd_>why not?
22:38<Trystan>you can configure it so you access the site but it doesnt ensure anything other than your code is correct
22:38<Trystan>because a virtual host relies on the domain name
22:38<sketchyd_>thats what i want to ensure - that my code and my database is restored properly
22:38<Trystan>that information is passed to the server, which then looks at your config and determines which site to give you
22:38<metap|pe>you could cut/paste the apropos apache confige entries into the default config under an alias
22:39-!-maladmin1 [~tom@] has joined #linode
22:39<maladmin1>hey all, am i live?
22:39<Trystan>sketchyd_: the host method would allow you to test it is restored properly while at the same time testing the virtual host as well
22:39<sketchyd_>im not understanding the virtual host thing
22:39<maladmin1>seems so, i just tried installing a new apache server but can't get ssl working correctly
22:39<maladmin1>nothing useful in the log files..
22:40<Trystan>whenever you try and access your website your computer does a lookup of where the site is hosted, say..
22:40<maladmin1>sketchyd_: are you talking apache virtual hosts?
22:40<sketchyd_>ok, trystan
22:40<Trystan>what hosts does, is tell the computer that it doesnt need to look at the DNS entry and tells it just to go to
22:40<Trystan>therefore you could test it, (not effecting your website or anything else) then remove the entry once you are happy and change the actual DNS entry
22:41<sketchyd_>ok - so at that point it would hit my linode
22:41<Trystan>yep, it would tell it just to go to the test site while the current one stays live
22:41<Trystan>then once you change the DNS
22:41<sketchyd_>right now, if i skip the first step, and simply type in, id just get that "it works" message
22:41<Trystan>it would change the linode to be the live site
22:41<sketchyd_>thats why im confused
22:41<maladmin1>lol i just spent the afternoon doing exacly that
22:41<sketchyd_>how do i do that?
22:42<sketchyd_>how do i get rid of that, i mean
22:42<maladmin1>sketchyd_: you need to change the index.html in /var/www/
22:42<Trystan>you could set the default home of your apache config to be where the files for your website are stored (the same place you told the virtual host)
22:42<Trystan>it would be reffered to as web root iirc
22:42<maladmin1>hmm get the feeling i'm halfway through a help here
22:42<Trystan>been a while since i fucked with apache
22:43<metap|pe>with virtual hosts, the ip address isn't important. just the fqdn
22:43<maladmin1>ip has to get the correct server
22:43<sketchyd_>to set this up, i followed:
22:43<maladmin1>what u trying to do sketchyd_
22:44<Trystan>test his website/virtual hosts
22:44<sketchyd_>thanks tryst
22:45<maladmin1>got a website setup already or starting from blank?
22:45<sketchyd_>im porting a website from another host
22:45<sketchyd_>ive copied all the files, restored the db
22:45<maladmin1>gallery website
22:45<sketchyd_>and i want to verify that it works before i move the domain
22:45<maladmin1>checked the permissions?
22:45<sketchyd_>dude, no
22:46<sketchyd_>it shows simply "it works"
22:46<sketchyd_>its not pointing to the right place
22:46<maladmin1>setup a file in sites-available?>
22:46<metap|pe>yes, it is pointing to the right place IF you configured it as a virtual host
22:47<Trystan>sketchyd_: I will put it to you another way
22:47<Trystan>you have one IP, and x number of sites
22:47<maladmin1>edit sites-avilable/default -- change DocumentRoot to point to the new folder
22:47-!-Sputnik7 [~Sputnik7@] has joined #linode
22:47<Trystan>not all of them can be accessed by accessing the IP, so virtual hosts are used
22:47<Trystan>they rely on the domain name to direct the user to the right place
22:47<sketchyd_>trystan - pm?
22:50<metap|pe>some horsies just won't drink the coolaid
22:50-!-bryanl [] has quit [Quit: Leaving...]
22:50<Trystan>why not!
22:56-!-cweilem [] has quit [Quit: You did what to who under what street light?]
22:58<tomg>Does anyone know if multicast between linodes is possible?
22:58<maladmin1>any hel with mod_ssl
22:58-!-memenode [] has quit [Remote host closed the connection]
23:02<bob2>maladmin1: you need to ask a question first...
23:02<bob2>maladmin1: which guide did you follow?
23:02-!-kassah [] has joined #linode
23:02<maladmin1>hmm.. all of them?
23:02<maladmin1>it really doesnt look hard
23:03<bob2>oh, sorry, I thought you were trying to set up ssl with your apache server
23:03<maladmin1>hey yup i am
23:03<maladmin1>installed and enabled mod_ssl
23:03<bob2>and then did what
23:03<maladmin1>created the certificate (also pointed to snakeoil)
23:04<maladmin1>i have default-ssl enabled
23:04<maladmin1>i think... checking
23:05<SelfishMan>tomg: multicast had best be killed on the network
23:07-!-robot [] has joined #linode
23:07<robot>hello peeoples
23:07<maladmin1>heh thx bob2, talking about it helps
23:07-!-robot is now known as nachtkriecher
23:07<nachtkriecher>anyone here done any firefox development?
23:10-!-Harry_Mudd [] has quit [Quit: WeeChat 0.3.0]
23:12-!-jacob [] has quit [Quit: Excess Fail]
23:12-!-ericoc [] has quit [Quit: leaving]
23:14-!-Hxm [] has joined #linode
23:16-!-argoe [~argoe@] has quit [Quit: argoe]
23:16-!-tomg [] has left #linode []
23:19-!-nachtkriecher [] has quit [Quit: leaving]
23:20-!-nachtkriecher [] has joined #linode
23:20<nachtkriecher>anyone here does firefox development?
23:21-!-maladmin1 [~tom@] has left #linode ['all fixed']
23:31-!-zack_ [] has quit [Quit: zack_]
23:33<Hxm>Are there any known performance issues going on now?
23:33-!-atula [] has quit [Quit: Leaving]
23:33-!-robot [] has joined #linode
23:34-!-robot is now known as nacht
23:41-!-Berto [] has quit [Remote host closed the connection]
23:41-!-Hxm [] has quit [Quit: Hxm]
23:43-!-jlorezon [] has joined #linode
23:44-!-jlorezon [] has quit [Remote host closed the connection]
23:44-!-zwit [] has joined #linode
23:44<amitz>Im gonna kill that moron who set his/her router without to any channel without consideration to my channel.
23:44<zwit>how's awake here?
23:44<amitz>I wish I can contact the owner somehow...
23:45<zwit>anyone know if recently deleted linodes are recoverable somehow?
23:45<straterra>if youre fast enough..
23:45<straterra>but not guaranteed
23:45<zwit>recently deleted disk images/config profiles to be precise..
23:46<@tasaro>zwit: is your account still active?
23:48<@tasaro>who's awake here?
23:48<straterra>I a,
23:48<@tasaro>was a joke ^
23:48<nacht>oh people
23:49<nacht>so why is it so difficult to make tabs (the character) a different size?
23:49-!-nachtkriecher [] has quit [Quit: Lost terminal]
23:49<Eman>sometimes they are 4 spaces, others they are 8 spaces
23:50<nacht>in firefox
23:50<bob2>why do you have tabs in things firefox is displaying
23:50<bob2>amitz: change yours
23:50<nacht>all kinds of reasons
23:50<nacht>no, 8 sucks
23:50<nacht>oh, to amitz, not to me
23:50<@tasaro>zwit: open a support ticket if you still have an account, otherwise email
23:51<nacht>i have code that gets displayed as text/plain
23:51<zwit>tasaro: yeah just opened a ticket
23:51<nacht>also i have textareas that edit code
23:52<zwit>hopefully it gets read soon enough
23:52<nacht>and they contain tabs
23:54<nacht>nobody here knows either?
23:54<Eman>amitz: set all your stuff to japan and use channel 14
23:55<@tasaro>zwit: taking a look
23:59-!-VS_ChanLog [] has left #linode [Rotating Logs]
23:59-!-VS_ChanLog [] has joined #linode
---Logclosed Tue Oct 27 00:00:51 2009