Back to Home / #linode / 2005 / 11 / Prev Day | Next Day
#linode IRC Logs for 2005-11-29

---Logopened Tue Nov 29 00:00:21 2005
00:04|-|dddd44 [dhb55@] has quit [Read error: Connection reset by peer]
00:15|-|spr [] has quit [Quit: Spoon!]
00:21|-|risto [] has quit [Ping timeout: 480 seconds]
00:24|-|cmantito [] has quit [Quit: Connection reset by peer? Oh yeah? RESET THIS PEER-BOY!]
00:25|-|cmantito [] has joined #linode
00:36|-|[|^__^|] [] has joined #linode
00:37|-|sec39 [] has joined #linode
00:48|-|risto [] has joined #linode
00:49<risto>was there a problem with host53 some time ago?
00:49<risto>my load suddenly went up and pings started dropping
00:54|-|Newsome [] has left #linode [Linux: Now with employee pricing!]
01:07|-|dddd44 [dhb55@] has joined #linode
01:25<caker>risto: not that I'm aware of
01:29<risto>ok..thanks caker... it happened about 0620 UTC == 01.20am Eestern. My load spiked up to 27 or so
01:31<risto>I'm suspecting it was some sort of network blurb that caused stuff to get backed up
01:32|-|cast [] has left #linode []
---Logclosed Tue Nov 29 01:44:30 2005
---Logopened Tue Nov 29 01:46:15 2005
01:46|-|mikegrb_ [] has joined #linode
01:46|-|Ekipa kanalu #linode: Wszystkich: 59 |-| +op [1] |-| +voice [0] |-| normalnych [58]
01:46|-|funkycow [] has quit [Ping timeout: 480 seconds]
01:46<tronix>ahh there it is.
01:46|-|cow [] has joined #linode
01:46|-|gpd [~gpd@] has quit [Ping timeout: 480 seconds]
01:46[~]tronix AOLs
01:46|-|mikegrb [] has quit [Ping timeout: 480 seconds]
01:46|-|tierra [] has quit [Ping timeout: 480 seconds]
01:46|-|efudd [] has quit [Ping timeout: 480 seconds]
01:46<tronix>maybe a network maint or something.
01:46<internat>lol yah for time outs
01:46|-|Kanal #linode zsynchronizowany w 34 sekundy
01:47<caker>(in PM or pastebin, please)
01:47|-|heidi [] has joined #linode
01:48<taupehat>caker: host15 load seems really high atm
01:48<caker>bad traceroutes, anybody?
01:48<tronix>looks ok for me now
01:48<tronix>came back after being out about 3-4 mins
01:48<caker>taupehat: host15 looks fine
01:48<caker>taupehat: which linode?
01:48<tronix>normally i'd say router maint but not supposed to be an issue with hsrp or vrrp
01:48<tronix>but knowing TP... *shrug*
01:49[~]tronix goes back to xml-style lurking &
01:50|-|gpd [~gpd@] has joined #linode
01:52|-|chris [] has joined #linode
01:52[~]caker shrugs
01:53[~]taupehat goes back to zzz-style snoring 2>&1
02:03<fo0bar>jesus, the UML version of finnix runs on 24MB of ram
02:27<internat>lol nice
02:27|-|internat [] has quit [Quit: This computer has gone to sleep]
02:36|-|ddsdsd [] has quit [Remote host closed the connection]
03:09|-|tierra [] has joined #linode
03:09|-|efudd_ [] has quit [Read error: Connection reset by peer]
03:09|-|tierra_ [] has quit [Read error: Connection reset by peer]
03:09|-|chris_ [] has joined #linode
03:09|-|chris [] has quit [Read error: Connection reset by peer]
03:10|-|efudd [] has joined #linode
03:10|-|heidi [] has quit [Read error: Connection reset by peer]
03:10|-|heidi [] has joined #linode
---Logclosed Tue Nov 29 03:10:24 2005
---Logopened Tue Nov 29 03:11:52 2005
03:11|-|mikegrb [] has joined #linode
03:11|-|Ekipa kanalu #linode: Wszystkich: 56 |-| +op [0] |-| +voice [0] |-| normalnych [56]
03:11|-|mode/#linode [+o mikegrb] by ChanServ
03:12<encode>mikegrb: whats going on with the network over at TP?
03:12|-|Kanal #linode zsynchronizowany w 29 sekundy
03:29<tronix>yeah just saw my linode drop out (network-wise) for 3rd time tonight
---Logclosed Tue Nov 29 03:32:42 2005
---Logopened Tue Nov 29 03:32:54 2005
03:32|-|mikegrb_ [] has joined #linode
03:32|-|Ekipa kanalu #linode: Wszystkich: 58 |-| +op [1] |-| +voice [0] |-| normalnych [57]
03:33|-|tierra [] has quit [Read error: Connection reset by peer]
03:33|-|efudd [] has quit [Read error: Connection reset by peer]
03:33|-|tierra [] has joined #linode
03:33|-|Kanal #linode zsynchronizowany w 28 sekundy
03:33|-|efudd [] has joined #linode
03:33|-|chris [] has joined #linode
03:33|-|chris_ [] has quit [Read error: Connection reset by peer]
03:33|-|heidi [] has quit [Read error: Connection reset by peer]
03:33|-|heidi [] has joined #linode
03:33<tronix>there... tp's back.
03:33<encode>for the moment
03:33<encode>how long will it last for though
03:34<tronix>heh point.
03:34|-|mikegrb [] has quit [Read error: Connection reset by peer]
03:35<tronix>outages seems to be about 4 mins or so... really does look
03:35<tronix>like they're probably loading new code on their routers,
03:35<tronix>if I had to guess
03:35<tronix>cause they typically take about that long to boot
03:36<tronix>just a little disappointing because if you have hsrp or vrrp
03:36<tronix>setup in first place, not supposed to be an outage at all.
03:36<tronix>with TP, one has to wonder if they set it up for Linode at all...
03:42<encode>interestingly, one of my ip addresses seems to be fairly stable, while the other has been dropping out
03:47<caker>it happened to me a little while ago, and it wasn't TP, it was AT&T, no where near their TP routes
03:48<caker>without traceroutes, don't blame it on TP
03:48<encode>where would you like a traceroute to?
03:48<caker>and yes, we have HSRP
03:48<caker>from you to your node's IP
03:48<encode>and how can i tell traceroute which interface to use?
03:48<encode>oh, ok
03:48<caker>(only when there's no/bad connectivity)
03:49<encode>i only notice on my irc connections, and since theres a 300 sec ping timeout, its kinda hard to know when the problem commences
03:49<encode>but i'll give it a try if it happens again
03:50<encode>why would it only affect one of my ip addresses though? it makes no sense
03:54|-|caritos [] has joined #linode
03:54<caritos>can anybody help assist me in why i'm not able to send mail with my postfix setup?
03:56<caker>encode: differnet subnets take different routes .. that makes sense to me
03:56<encode>i spose so
04:04|-|internat [] has joined #linode
04:07|-|thoth39 [] has joined #linode
04:15|-|andrewjw [] has joined #linode
04:35|-|dddd44 [dhb55@] has quit [Read error: Connection reset by peer]
05:12|-|caritos [] has quit [Remote host closed the connection]
05:39|-|lhurgoyf [~can@] has joined #linode
05:40<lhurgoyf>hi linbot
05:40<linbot>high on life, lhurgoyf
05:44|-|lhurgoyf [~can@] has left #linode []
05:45|-|dddd44 [dhb55@] has joined #linode
05:47|-|dddd44 [dhb55@] has quit [Read error: Connection reset by peer]
05:50|-|Shaun2222 [] has joined #linode
05:57|-|Shaun [] has quit [Ping timeout: 480 seconds]
05:58|-|Shaun2222 changed nick to Shaun
06:29|-|Netsplit <-> quits: darkbeholder, internat
06:34|-|Netsplit over, joins: internat
06:44|-|FireSlash [] has joined #linode
07:03|-|FireSlash [] has quit [Read error: Connection reset by peer]
07:12|-|internat [] has quit [Quit: This computer has gone to sleep]
07:16|-|linville [] has joined #linode
07:24|-|internat [] has joined #linode
07:24|-|internat [] has quit [Quit: ]
07:25|-|internat [] has joined #linode
08:06|-|darkbeholder [] has joined #linode
08:46|-|thoth39 [] has quit [Quit: Leaving]
08:48|-|emcnabb [] has joined #linode
09:05|-|Redgore [~Redgore@] has joined #linode
09:07|-|Newsome [] has joined #linode
09:10|-|mikegrb_ Your nick is now mikegrb_|
09:43|-|Newsome [] has quit [Quit: Linux: Now with employee pricing!]
09:55|-|Bducharme [~Bducharme@] has joined #linode
09:57<Bducharme>are there any servers down?
09:57<kvandivo>undoubtably, yes
09:57<Bducharme>I am trying to figure out a problem with li4-29
09:58<Bducharme>Even after reboot still unable to connect through ssh.. ;(
09:58<kvandivo>connect via lish and debug it
10:00<Bducharme>if I can remember how to do that.. there's info in the control panel area if I remember correctly
10:01<npmr>in the "Host Summary" box on the control panel it says which host your linode is on
10:01<npmr>ssh to that host using your linode username
10:02<Bducharme>example right?
10:03<Bducharme>got it... thanks
10:04|-|bendy24 [] has quit [Ping timeout: 480 seconds]
10:06<mikegrb_|>Bducharme: you were over 200 mb into swap before rebooting, I'd look into your ram usage
10:07<mikegrb_|>after rebooting you are already 30 mb in
10:07<mikegrb_|>100 now
10:08<Bducharme>wonder what the hell is going on...
10:08<Bducharme>mysql crashed on me yesterday and I lost a few tables
10:09<mikegrb_|>you probably ran out of ram + swap and the oom killer killed mysql to free ram
10:13<Bducharme>any ideas on how to bring it up with mysql stopped?
10:13|-|bendy24_ [] has joined #linode
10:15|-|spr [] has joined #linode
10:18|-|spr [] has quit [Quit: ]
10:23|-|dddd44 [dhb55@] has joined #linode
10:24|-|besonen_ [] has quit [Quit: Leaving]
10:27|-|smitty [] has joined #linode
10:27<smitty>anyone there?
10:27<smitty>does anyone have experience running "ha" on a linode?
10:27<smitty>ha = High Avalibility
10:28<linbot>New news from forums: High availability in General Discussion <>
10:28<smitty>that is my post
10:29<Redgore>linbot is a bot
10:29<linbot>hmm... is a bot is a sighup, Redgore
10:29<smitty>at least someone is here
10:30|-|smitty [] has quit [Quit: ]
10:39|-|Newsome [~sorenson@] has joined #linode
10:41|-|besonen [] has joined #linode
10:57|-|andrewjw [] has quit [Quit: using sirc version 2.211+KSIRC/1.3.12]
10:59|-|mikegrb_| Your nick is now mikegrb
11:20|-|Bducharme [~Bducharme@] has quit [Quit: Bducharme]
11:57|-|spr [] has joined #linode
12:13|-|J[SS] [] has quit [Read error: Connection reset by peer]
12:14|-|J[SS] [] has joined #linode
12:29|-|Jeremy [] has joined #linode
12:30|-|Jeremy [] has quit [Quit: ]
12:46|-|emcnabb [] has quit [Read error: Connection reset by peer]
12:49|-|lhurgoyf [~can@] has joined #linode
12:55|-|darkbeholder [] has quit [Remote host closed the connection]
13:09|-|emcnabb [] has joined #linode
13:09|-|emcnabb [] has quit [Quit: ]
13:15|-|Redgore [~Redgore@] has quit [Quit: A geek without purpose - | SMDC-Network IRC -]
13:17|-|Redgore [~Redgore@] has joined #linode
13:20|-|Electric1lf changed nick to ElectricElf
13:26|-|Newsome [~sorenson@] has quit [Quit: Linux: Now with special employee pricing!]
13:35|-|UNLEASHED [] has joined #linode
13:35<UNLEASHED>what is linode
13:36<caker>read /topic
13:36<UNLEASHED>do they do dedicated servers?
13:36<anderiv>UNLEASHED: find out for yourself
13:36<UNLEASHED>if so cool i need
13:36<lhurgoyf>linode is a word which has 6 letters.
13:37<UNLEASHED>all i want to know is if they want the 4000 dallors i am going to spend on servers if not i will look elsewhere
13:37<lhurgoyf>(what? linode) --> (what? word) (what? letter)
13:38<UNLEASHED>ok nevermind
13:38[~]caker likes dallors
13:39<anderiv>UNLEASHED: if you ask my, linode provides an incredible service, but like anything, it may be the right decision for you and it might not. Read up on the website and see what you think.
13:39<anderiv>**** if you ask *me*, rather
13:39<lhurgoyf>dal is for "leaf" and lor is for "cheese" in turkish.
13:40<lhurgoyf>a type of chese
13:40[~]anderiv likes leaf cheese
13:41<lhurgoyf>i wouldn't buy a linode if bandwidth is cheaper in turkey
13:41<kvandivo>i wouldn't care what bandwidth was, only what bandwidth is
13:41<lhurgoyf>you pay for excess bandwith, starting after 1GB !
13:42<kvandivo>in soviet russia, bandwidth pays for you
13:42<lhurgoyf>a friend told me that you pay more for traffic comingfrom the world. intercountry traffic is cheap
13:43<caker>kvandivo: :)
13:44<lhurgoyf>i am monitoring the logsize of gaim via cacti. it's fun! I suggest anyone having a little spare time for setting up
13:45<anderiv>lhurgoyf: nerd ;-)
13:47<lhurgoyf>anderiv: if i find a module which provides a proc interface for keycounts...
13:48<caker>grep keyboard /proc/interrupts
13:48<anderiv>ooh - look at the hotshot ;-)
13:50<anderiv>lhurgoyf: where in Turkey are you?
13:50<lhurgoyf>timer i8042 cascade ATI IXP, yenta ohci1394, ndiswrapper rtc acpi ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3, yenta radeon@pci:0000:01:05.0 i8042 ide0 ide1 only. no keyb*
13:51<lhurgoyf>seleck Can B... from the list
13:51<lhurgoyf>that is the exact location
13:54<anderiv>did y'all see that firefox-1.5 was released?
13:54<anderiv>it's on the ftp servers, but hasn't been updated yet.
13:55<lhurgoyf>anderiv: they are probably waiting for mirrors
13:55<anderiv>lhurgoyf: yah probably
14:03<lhurgoyf>anderiv: i submitted to reddit :P
14:04<anderiv>lhurgoyf: I'm not familiar w/ reddit
14:22|-|chris_ [] has joined #linode
14:22|-|chris_ [] has quit [Quit: ]
14:35<anderiv>caker: so you've recovered from all those issues w/ borked hard drives last week?
14:36<caker>anderiv: yup. Only one host was unplanned
14:37<anderiv>caker: ahh okay - for some reason I thought there were two.
14:37<caker>two planned, one unplanned
14:37<caker>that reminds me
14:37[~]caker orders more drives
14:37<anderiv>caker: how many customers do you have? Or do you not divulge that info...
14:37<caker>Just ordered our first AMD box today
14:38<lhurgoyf>anderiv you can calculate that
14:38<caker>anderiv: rather not .. but, do the math .. ~50 hosts online (rest have dead drives or are dev boxes)
14:38<anderiv>caker: right on
14:39<anderiv>caker: AMD64 or 32?
14:39<caker> <--
14:40<caker>8G ram (2G chips were cheaper than 2 1G chips)
14:40<caker>AMD ... 246 procs
14:40<caker>that was the sweet spot in pricing for AMD procs
14:40<efudd>2 1gig is better than 1 2gig tho
14:40<anderiv>do you run those raid5?
14:41<caker>some of them are crazy expensive.. dual core, FX (whatever the hell that is), etc
14:41<efudd>FX is badass
14:41<caker>anderiv: raid1
14:41[~]efudd nod wildly
14:41<caker>what does it do?
14:42<efudd>caker, that's the moniker for AMD's top end processors
14:42<tsi>i wish monster would make their website more sfw
14:42<efudd>their FX-57 or whatever completely owns in most(all?) benchmarks
14:42<lhurgoyf>Athlon FX is apparently going to be AMD's official name for the new Athlon XP Thorton core!
14:43<caker>AMD Athlon. 64 FX-57 ~ $1000 bucks
14:43<efudd>yup. it isn't cheap.
14:45<caker>I'm hoping (from what I've read) that a box with this spec will not only outperform my last intel host spec (dual 3.2 Xeon, yaddayadda) but save a decent amount of power too
14:45<JasonF>caker: zomg! Finally amd64!
14:45[~]JasonF does a happydance
14:45<anderiv>caker: do you pay per kWh?
14:46<caker>anderiv: no, but power is a limited resource ... less power/box means I can utilize more of the cabinet space
14:46<anderiv>caker: ahh - righto
14:46<caker>TP limits me a two 20A circuits per cabinet, and HE is (freaking insane) with only a single 15A circuit
14:46|-|Newsome [] has joined #linode
14:47<caker>nope, 120
14:47<anderiv>ugh - a single 15A.
14:47<anderiv>caker: that's rediculous.
14:47<fo0bar>caker: HE takes its model from Exodus, which my old datacenter/employer did as well. thankfully we realized 20A/rack was stupid early on
14:47<fo0bar>... and then we went out of business
14:47<caker>Years ago .. when I first ran up agains the power limits at HE .. the chick was like "Well, plenty of our customers pack their cabinets with 40 servers on 15A"
14:48<caker>um .. 15A / 40 servers is about 1/3 of an amp .. about what it takes to run an IDLE hard drive now a days
14:48[~]anderiv just ran a 20 amp circuit to his *home* datacenter. :-)
14:49[~]fo0bar runs out to the car to grab the Kill-A-Watt
14:49<fo0bar>I'm curious to see what my current supermicro generation is drawing
14:49<caker>oh, need of those
14:49[~]caker wonders if Radioshack has something I can use
14:50<tsi>as long as it's not something you need
14:50<caker>mikegrb: the inductance kind, or whatever it's called, that you just clamp over the wire?
14:50<mikegrb>they have a consumer friendly thing that can even display $/hr
14:50<mikegrb>they also have an ammeter probe that clamps
14:50<mikegrb>comes with a special pass through plug with a loop of wire
14:50<tsi>but you'll be much happier with the one at home depot
14:51<caker>tsi: thanks for the advice
14:51<mikegrb>if you just put the clamp around the single cable, the two conductors cancel each other out
14:51<caker>tsi: model number or something to inquire about while there?
14:51<anderiv>mikegrb: you just want the hot wire, I assume?
14:52<mikegrb>well, either wire
14:52<mikegrb>the clamps from radio shack come with a passthrough plug
14:52<mikegrb>it plugs into the outlet and the device plugs into it
14:52<tsi>wiring aisle, there'll be a dozen models that look like rubber-coated pliers with a pda attached
14:52|-|spr [] has quit [Quit: Spoon!]
14:52<caker>tsi: mkay
14:52<mikegrb>it has a loop of wire sticking out the side to test
14:53<anderiv>mikegrb: that's slick.
14:53[~]tsi avoids going into detail on phazor notation and imainary component electric field
14:54[~]mikegrb adjusts tsi's real power
14:54[~]Newsome seconds fo0bar's suggestion of getting a Kill-A-Watt
14:54<tsi>Watts Up is a brand of consumery plug-in meters
14:55<tsi>the downside being you have to disrupt power supply to use it
14:55<tsi>(i.e. dealbreaker)
14:55<mikegrb>tsi: you would need to for any meter
14:55<mikegrb>tsi: unless you already have the conductors seperated
14:55<mikegrb>or don't mind slicing the outer jacket on the power line while it's running
14:56<tsi>that ammeter scares me
14:57<mikegrb>hmm, it doesn't mention the little plug thingie in the manual so don't know if it comes with it
14:58<tsi>yeah and those built in leads don't look rated for more than about 1a
14:59<fo0bar>ok, the only (supermicro) machine I had in back that I could unplug was a P4 3.2ghz ht/1GB/2x 160GB 3ware
14:59<tsi>i forget what i went to to look up, but i've decided we all need 400psi 80 gallon compressors
14:59<mikegrb>500 V
14:59<mikegrb>DC or RMS AC
14:59<mikegrb>tsi: yes, get me one for christmas
14:59<fo0bar>POST: 1.7A, idle: .9A, thrashing CPUs and disk: 1.55A
15:00<tsi>!calc 1.8*120
15:00<linbot>tsi: 1.8 * 120 = 216
15:00<mikegrb>tsi: tried to talk heidi into a brad nailer, finish nailer, compressor kit... ended up with a battery powered brad nailer
15:00<tsi>*almost* the same.... :\
15:01<mikegrb>I /reallly/ wanted the flexability of having the compressor
15:01<fo0bar>caker: radioshack USED to sell the kill-a-watt, but not anymore
15:01<tsi>dear women, "torque" is to us what "sale" is to you. love, men.
15:01<mikegrb>air tools are pretty damn cheap since they externalise the power source
15:01<efudd>oh no. my name just got recorded as "hashnarishamanajanba-you-can-call-me-bob" for about half of the company on a conference call that had name-announce turned on.
15:01<caker>fo0bar: ok stupid questions time: P4 as in P4, or p4 as in Xeon? Also, what model SM is that?
15:02<fo0bar>caker: regular P4, we don't deal with xeons much here. 5014C-MT
15:02<fo0bar>don't use my calculations for anything ;)
15:02|-|futhin [] has quit [Remote host closed the connection]
15:02|-|Galateja [] has joined #linode
15:02|-|futhin [] has joined #linode
15:02<Newsome>also, is that per-PS, or total?
15:02<caker>fo0bar: ok, our Xeon boxes use quite a bit more power than that
15:02<fo0bar>caker: I'd imagine
15:02<fo0bar> <-- caker
15:02<fo0bar>I'm not sure if that's available in-store
15:03<fo0bar>but that's what I have
15:03<Newsome>Catalog/ Internet Only :(
15:03<fo0bar>Newsome: ahh, didn't see that
15:03<tsi>awww it's so cute
15:03<fo0bar>it'll also do kWh, which is useful
15:05<tsi>so can bc
15:05<caker>2.26A RMS average, 3.06A peak average (across 6 machines) <-- dual 3.2 xeon (6014H-T)
15:05<caker>so, I get 6 boxes per 20A cicruit ...
15:05<caker>12 per cabinet .. what a waste
15:05<tsi>even that sounds low
15:06<tsi>(current peak, not server density)
15:07<tsi>wait, per machine or for all six?
15:07<tsi>durrrrrrr i'm illiterate
15:07<caker>True RMS Current: 13.6 Amps
15:07<caker>Maximum Detected: 18.4 Amps
15:07<caker>^-- totals
15:07|-|Galateja [] has quit [Remote host closed the connection]
15:08<tsi>yeah that sounds in line with my experience
15:12|-|futhin [] has quit [Ping timeout: 480 seconds]
15:13|-|futhin [] has joined #linode
15:14|-|jimcooncat [] has joined #linode
15:15<jimcooncat>confused about postfix (not the first time). How do I block a small list of IP addresses? with cidr_table?
15:22|-|rafx [] has joined #linode
15:23|-|alan [] has joined #linode
15:35|-|jimcooncat [] has quit [Quit: Client Exiting]
15:36<gpd>greetings all... good thanksgiving?
15:36<anderiv>gpd: yup!
15:36<gpd>good news :)
15:36<gpd>my mailman spam list is huge... which prompts investigation...
15:36<lhurgoyf>i havea strangefeeling that cacti doesn't graphs oidswhich returns "string" which contains a number
15:38<gpd>does anyone have a spam filter regex for mailman?
15:39<gpd>I suppose I could set auto reject...
15:43<anderiv>gpd: this is what I did with my thanksgiving --->
15:44<anderiv>gpd: non-typical thanksgiving for me, but good nonetheless
15:44<gpd>nice work!
15:44<anderiv>gpd: hehe - thanks.
15:45<gpd>I stuck with the more traditional eating and drinking too much... oh and too much shopping...
15:46<anderiv>gpd: thankfully, I haven't set foot in a retail establishment in weeks :-)
15:47<kvandivo>9 pounds 4 ounces! ouch
15:48<anderiv>kvandivo: hehe - my lovely wife did very well.
15:49<kvandivo>she wasn't condemning you to a fiery death in the lowest depths of hell during the ordeal?
15:49<anderiv>kvandivo: ha!
15:50<anderiv>kvandivo: nope. of course the epidural helped w/ that.
15:50<tierra>congratz anderiv
15:50<anderiv>tierra: tanks
15:56|-|thoth39 [~hm@] has joined #linode
16:11<caker>anderiv: congrats!
16:11<anderiv>caker: thank ya, sir.
16:13|-|linville [] has quit [Quit: Leaving]
16:14|-|Jeremy [] has joined #linode
16:14|-|Jeremy [] has quit [Quit: ]
16:18|-|lhurgoyf [~can@] has quit [Quit:]
16:39<mikegrb>2 pk socks + 3 pair panties 2.49 shipped
16:39<mikegrb>holy cow
16:39<caker>mikegrb: just for you
16:40[~]mikegrb orders 10
16:40<mikegrb>a whole month worth of undies for $20!
16:41<mikegrb> Dell 2005FPW Ultrasharp Widescreen 20" LCD Monitor $334 <-- expired :<
16:41<kvandivo>does that mean 2 sets, or are you ordering 70 pieces of clothing?
16:43<mikegrb>10 sets
16:45<fo0bar>mikegrb: so, dell coupons expire "12/14, or whenever we choose"?
16:45<npmr>caker, are the 2.6.12.* linode kernels built from vanilla source?
16:45<caker>npmr: it's all on this page:
16:46|-|Redgore [~Redgore@] has quit [Quit: A geek without purpose - | SMDC-Network IRC -]
16:46<caker>npmr: why do you ask?
16:47<npmr>i'm building myself a new uml kernel for the host in the basement, looking for sources to start with
16:47<gpd>why not use Xen instead...
16:47[~]gpd ducks
16:47<npmr>Linux egg 2.6.4-uml1-1um #242 Sun May 9 22:03:00 EDT 2004 i686 GNU/Linux
16:48<fo0bar>npmr: 2.6.14 mainline + -bs patches work well
16:48[~]anderiv reminds himself that he really needs to play w/ xen here soon.
16:48<caker>npmr: recently I built 2.6.14 from vanilla just fine, for what it's worth...
16:48<fo0bar>actually 2.6.14 mainline works well too
16:48<npmr>how's the nptl stuff?
16:48<npmr>still broken?
16:49<caker>There are patches on Jeff's site, but he said (today) it likely won't go in until 2.6.16
16:49<gpd>caker: are you skipping 2.6.14 and waiting for 2.6.16 for linode hosts?
16:50<caker>gpd: I have hosts running 2.6.14 now
16:50<gpd>how about for us mortals?
16:50<caker>ahh, no I should have a .14 out Real Soon Now
16:51<caker>if Linus and friends beat me to it, a .15
16:51<gpd>.15 is in Dapper... but aren't the odd ones 'unstable'?
16:51<fo0bar>caker: 2.6.15 is still probably a few weeks out
16:52<caker>No, odd minors are unstable "major.minor.?revision"
16:52<npmr>17:52:00 (346.23 KB/s) - `linux-' saved [39186327/39186327]
16:52<npmr>dang cable's fast
16:53|-|vaxen [] has joined #linode
16:53<fo0bar>and one day, we will be blessed with the mythical 2.7... yeah right
16:53<gpd>I thought that was the old rule... and the new rule was any 2.6.odd was more unstable than 2.6.even ???
16:53<caker>gpd: nope
16:53<gpd>probably wrong... :(
16:53<fo0bar>gpd: no, 2.6.NN is more stable than 2.6.NN-rcN :)
16:54<linbot>gpd: *click*
16:54<vaxen>emm..there is no wordpress deb for debian
16:55<internat>yes there is
16:55<caker>fo0bar: I bet in time for Christmas
16:55<fo0bar>caker: probably
16:55<gpd>there is one in ubuntu... from universe - so should be
16:55<vaxen>ohya, in unstable
16:56<internat>wordpress - an award winning weblog manager
16:56<internat>its in testing
16:56[~]anderiv is going to do his first ubuntu install this evening. I've heard good things...
16:56<fo0bar>anderiv: it will brew you coffee in the morning
16:56<gpd>ubuntu++ - now on my desktop for 6mo... and only rebooted into XP 3 times...
16:57<caker> 17:56:44 up 575 days, 2:25, 1131 users, load average: 2.25, 2.47, 2.15
16:57<caker>a few others, too
16:57<vaxen>i'm gonna try update my gentoo again, lets hope there wont be another kernel panic
16:57<anderiv>fo0bar: hehe
16:57<caker> <-- fo0bar's
16:57<anderiv>gpd: I'm a hardcore gentoo evangelist for the server end of things (headless installs), but I'd never want to run it on a desktop. That's why I'm trying ubuntu.
16:58<gpd>vaxen: that wordpress in unstable is v. 1.5.2... screw that and install v2 it by hand
16:58<anderiv>gpd: have you tried wordpress v2 yet?
16:58<caker>anderiv: not to start a distrowar or anything, but. In your opinion, what's the single biggest benefit to Gentoo on a server vs a binary distro?
16:58<gpd>anderiv: no :(
16:59|-|artifex [] has joined #linode
16:59<caker>hello arti
16:59<artifex>hi :-)
16:59<anderiv>caker: hrm - good question. portage is great, but I know that can be installed on other distros now. I guess I've just gotten used to the "gentoo-way-of-doing-things".
16:59<caker>anderiv: fair enough
16:59<anderiv>caker: and it's been quite stable for me in a production env.
17:00<vaxen>gpd: I would, but i can't be bothered to check for latest versions all the time
17:00<fo0bar>fo0bar's setup: gentoo on the laptop and media center, ubuntu on the desktop, debian on the server
17:01<anderiv>heh - I'm pulling the ubuntu iso down from the University of Minnesota's mirror at ~1.5Mbit. Niiiice.
17:02<vaxen>whats the learning curve for debian?
17:02<fo0bar>lapto because gentoo's packages are usually newer than other distros, and it's very easy to compile 3rd-party kernel modules in the portage tree (ndiswrapper, ati-drivers, alsa-drivers, etc)
17:02<caker>vaxen: apt-get install foo. done
17:02<anderiv>caker: I'm not naive to all of the disadvantages of gentoo by any means.
17:02<vaxen>I just wanna get my apache+wordpress up and running asap
17:02<fo0bar>media center because gentoo is the only distro I've found that handles lirc packages in a sane way
17:02<vaxen>can you mix unstable and stable packages in debian?
17:02<anderiv>caker: I just haven't been stung by any of them yet.
17:03<anderiv>*** knocks on wood.
17:03<gpd>vaxen: ubuntu breezy + apt-get install wordpress (job done)
17:03<kvandivo>vaxen: depends on the phase of the moon
17:03|-|artifex changed nick to Artifex
17:03<anderiv>kvandivo: hah
17:03<anderiv>kvandivo: well put
17:03<vaxen>kvandivo: ?
17:04<kvandivo>vaxen: if you're extremely lucky you can mix.. otherwise you'll have dependency problems
17:04[~]gpd adds to localhost redirect in /etc/hosts :<
17:04<vaxen>but i can do that in gentoo :) no flames
17:05<anderiv>gpd: don't like digg?
17:05<efudd>what? doesn't suck.
17:05<gpd>no... I just find myself less productive at work!
17:05<anderiv>gpd: ha
17:05<gpd>... note to self... stop using irc...
17:06<vaxen>do you guys run debian testing or stable?
17:06<gpd>stable atm... no reason to switch on the linode so far...
17:06<gpd>that is another nice thing about ubuntu... backports :)
17:11<fo0bar>debian's at that point right now where separate backports usually aren't needed (IE, you can usually grab a .deb from testing and throw it on stable)
17:11<fo0bar>but etch is slowly diverging
17:11<npmr>Filesystem in Userspace support (FUSE_FS) [N/y/?] (NEW)
17:11<fo0bar>I keep hearin about FUSE... sounds cool
17:11<fo0bar>but I haven't played with it yet
17:12<anderiv>fo0bar: it's pretty slick...I've just played w/ sshfs a bit, and it worked as advertized.
17:14<tierra>I've been curious to play with it at home for the single filesystem "Offline Media Content Database" listed on FUSE
17:14<tierra>but have been finding it hard to find much info on it
17:15<tierra>but it sounds like something I could definately take advantage of
17:18<alan>is there that much use for it (other than what you can do with ssh+mc+shell-link and/or scp
17:21<tierra>well, there potentially could be depending on it's features, but I can't even find info on what it supports
17:21<tierra>given I haven't had time to look at the code itself for some of the other frontends
17:23<tierra>on that note, I haven't even been able to find the actual FUSE code for it, just the php and cli frontends along with some sql
17:24<tierra>oh, sorry, were you talking about sshfs?
17:24<alan>tierra: yes
17:29<gpd>I am disturbed by the $60 price tag for 360 games...
17:30|-|spr [] has joined #linode
17:38|-|iggy_ changed nick to iggy
17:40<mikegrb>gpd: and not the $50 for xbox games?
17:41<mikegrb>I agree, $60 is a lot for games (for me anyway, people who play 10 hours a day certainly would get their $60 worth), but I expected it to be higher then that
17:41<gpd>I think you are right - but it doesn't mean I have to like it
17:42<gpd>I would rather have a subscription... any game for 20 hours a month... for example
17:42[~]mikegrb agress
17:43[~]mikegrb goes to eat
18:05|-|emcnabb [] has joined #linode
18:24|-|darkbeholder [] has joined #linode
18:30<alan>hmm tried out sshfs but hit the bug that /dev/fuse must be manually created. iirc creating that file in dev is not as simple as mkdir
18:31<alan>oh yeah thats it thanks
18:31<caker>Google says to: mknod /dev/fuse -m 0666 c 10 229
18:32<alan>great. that looks familiar, i had to do it for some other thing once, something with cd i think
18:34<caker>Quote of the day: When you're in a hole, quit digging
18:36<alan>bah, fuse device not found and modprobe would load it
18:36<caker>obvious question: this isn't your Linode, correct?
18:37<alan>mine was ssh target
18:37<caker>Linodes don't do modules
18:38|-|UNLEASHED [] has quit [Remote host closed the connection]
18:38<alan>actually the src is a xen
18:38<caker>ahh .. build the module with ARCH=xen, then?
18:39<alan>actually just playin, but it might have been useful with the linode as the src machine ... .
18:40<tierra>I've been using shfs for a while now and have been pretty happy with it (minus some fstab ssh key issues)
18:40<tierra>shfs != sshfs btw
18:42<alan>tierra: will work on a uml or xen?
18:43<tierra>not with your linode as the machine your mounting a remote directory onto (as shfs has to be built as a kernel module)
18:44<tierra>but you could mount a directory on your linode on a remote machine with shfs loaded
18:45<tierra>but given a UML with kernel module loading support, yes
18:56|-|spr [] has quit [Ping timeout: 480 seconds]
18:57|-|thoth39 [~hm@] has quit [Quit: Leaving]
19:03|-|dsp_ [] has quit [Quit: meh]
19:06<iggy>anybody heard anything about strange signals being thrown in certain programs?
19:06|-|dddd44 [dhb55@] has quit [Read error: Connection reset by peer]
19:10[~]mikegrb throws iggy strangely
19:19<internat>iggy, what os
19:19<internat>and what kernel
19:24|-|bendy24_ [] has quit [Read error: Connection reset by peer]
19:24|-|bendy24 [] has joined #linode
19:24|-|emcnabb [] has quit [Ping timeout: 480 seconds]
19:32|-|darkbeholder [] has quit [Quit: brb]
19:37|-|darkbeholder [] has joined #linode
19:39|-|spr [] has joined #linode
19:51[~]caker checks out
19:51<caker>pretty cool
19:52<caker>!errno 9
19:52<linbot>caker: EBADF (#9): Bad file descriptor
20:09|-|internat [] has quit [Quit: This computer has gone to sleep]
20:23<linbot>New news from forums: did uncle xen skip town? in Xen Testing <>
20:40<JasonF>!errno 7
20:40<linbot>JasonF: E2BIG (#7): Argument list too long
20:40<JasonF>!errno 733
20:40<linbot>JasonF: (unknown) (#733): Unknown error 733
20:40<JasonF>!errno 73
20:40<linbot>JasonF: EDOTDOT (#73): RFS specific error
20:40<iggy>!errno 6
20:40<linbot>iggy: ENXIO (#6): Device not configured
20:49<JasonF>!errno 1
20:49<linbot>JasonF: EPERM (#1): Operation not permitted
20:49<JasonF>I got that error from my girlfriend the other day :(
21:03|-|bendy24 [] has quit [Read error: Connection reset by peer]
21:03|-|bendy24 [] has joined #linode
21:12|-|tag [] has joined #linode
21:13<tag>compressed file systems?
21:13<npmr>cramfs comes to mind
21:14<tag>I wonder if it's going to painfully slow, I was thinking about repartitioning my linode or grabbing another chunk of disk space, and throwing a compressed filesystem on it for my respository
21:14<npmr>also, cloop.... compressed block device
21:14<npmr>although that would require loading a kernel module, which you can't do with linode
21:14<tag>can we get compressed file systems in linodes?
21:15<npmr>[inkblot@dorothy:~]$ zgrep CRAM /proc/config.gz
21:16<mikegrb>cramfs is read-only though, right?
21:16<npmr>i believe so, yes
21:16<npmr>as is romfs for sure
21:16<npmr>and probably zisofs as well
21:17<npmr>tag, is compression really going to solve your problem anyway?
21:18<npmr>i'm a little surprized that there's are no compression targets for device mapper
21:18<npmr>dm-gz, dm-bz, dm-zip, (etc)
21:19<tag>npmr: or more money :-)
21:19<npmr>or less data
21:19<tag>less data would be bad.
21:20<tag>because this is my repository, guy!
21:20<tag>this is my code, the more revision history I have the better.
21:20<npmr>is it necessary to keep it on the linode?
21:20<npmr>could you host the repository on some other system?
21:20<tag>I purchased the linode specifically for this purpose.
21:20<npmr>i see
21:21<tag>I don't want to pay for a full dedicated server just to have my repository and personal website
21:21<Newsome>pester caker about getting a newer kernel with FUSE support
21:21<npmr>how about dsl and a server in your coat closet?
21:21<tag>but I don't want to have my repository and website in a horribly unreliable my dads closet, where it was years ago, or some employers machine I've had alot of issues with that over the years ;-)
21:23<npmr>surely you must have a sympathetic maths teacher or something who could spare an outlet, a port, and a couple cubic feet
21:23<caker>squashfs is another
21:23<npmr>caker, i've got a kernel running dog slow
21:24<npmr>thirty minutes into the boot sequence and klogd is just starting
21:24<caker>npmr: geesh
21:25<npmr>caker, have you seen anything like this?
21:26<tag>npmr: but then I'd have to get rid of my linode, are you really trying to chase away business in here?
21:26<npmr>tag, you could run both
21:26<npmr>tag, i did
21:26<tag>what would be the point of that.
21:26<caker>npmr: no... the finnix recovery kernel is, metehinks .. it's a cut-down .config, but I didn't see slowness like that, or any slowdown at all
21:27<caker>tag: what's the problem? Need more space?
21:27<npmr>caker, kernel started up fine, it was when it hit userspace that it slowed to a crawl
21:27<caker>npmr: any other patches?
21:27<npmr>caker, vanilla
21:27<npmr>no patches at all
21:27<caker>npmr: running in skas3 mode?
21:27<Newsome>tag: what's the revision-control system?
21:27<caker>and it actually detected skas3?
21:28<npmr>2.6.8 debian sources plus kernel-patch-skas
21:28<npmr>it did
21:28<tag>caker: yeah...trying to figure out how to optimize my linode for low traffic areas of the file system
21:28<tag>Newsome: subversion
21:28<caker>tag: ok, depending on the distro there are a few places you could rm to get some space back
21:28[~]Newsome wouldn't use a kernel as old as 2.6.8 (for host or UML)
21:29<caker>for instance: /usr/share/doc
21:29<caker># du -sh /usr/share/doc/
21:29<caker>120M /usr/share/doc
21:29<caker>^-- RH9
21:29<tag>oh yeah
21:29<Newsome>make sure logrotate compresses old logfiles
21:30<tag>I am with bzip
21:30<caker>apt-get autoclean
21:31<iggy>manpages, infopages, if you don't need them
21:31<caker>anything in /boot and /lib/modules (if they somehow got populated)
21:31<fo0bar>npmr: I have a .deb of with skas3 applied
21:31<npmr>Newsome, was running 2.6.4 until yesterday
21:32<Newsome>wow :)
21:32<mikegrb>also locals
21:32<mikegrb>localpurge > *
21:32<npmr>fo0bar, build #242?
21:32<fo0bar>npmr: no, I made it myself
21:32<mikegrb>localepurge - Automagically remove unnecessary locale data
21:32<caker>oh yeah
21:32<caker>that's a good one
21:33<mikegrb>it's in debian probably in ubuntu universe or metaverse
21:33<caker>~50M on RH9 for locales I don't need
21:33<npmr>fo0bar, you know how to set to build number, right?
21:34[~]caker wonders when .version overflows
21:35<fo0bar>npmr: not off the top of my head
21:35<vaxen>apt-get will remove dependencies of the package you are going to remove, what happens when those dependencies are dependent for other packages?
21:36<npmr>fo0bar, echo 241 > .version
21:36<vaxen>i know for sure gentoo portage don't handle that very well, wondering if apt does it better
21:36<npmr>apt knows better
21:36<mikegrb>does apt removed unused dependancies? I know aptitude does
21:37<vaxen>isnt aptitude a frontend to apt?
21:37<mikegrb>front-end to dpkg
21:37<vaxen>i guess gentoo portage got a long way to go to catchup
21:37<mikegrb>as is apt
21:37<npmr>apt-get does not remove unused dependencies
21:37<npmr>aptitude is a front end to apt
21:37<mikegrb>or deselct
21:38<npmr>as is apt-get
21:38<tag>word up
21:38<tag>I have space, I was just trying to think ahead 'cause now that I lost my full time employment I'm doing alot of consulting work and my respository is thus much busier ;-)
21:39<encode>tag: you could keep historical data offsite
21:39<encode>the backup / offsite service i use is $5 USD / month per 10GB
21:39<fo0bar>mikegrb: thanks for localepurge, I had been doing it by hand until now
21:40<encode>i haven't really had any issues with them, had fast response times to queries
21:40<tag>I don't know if its possible to shave off revision history in subversion...hurm
21:40<caker>npmr: want to try that finnix kernel I built?
21:40<tag>encode: what service is that though, I'd like to send incremental dumps to it
21:40<Newsome>tag: pester caker for FUSE, and run gmailfs!
21:40<encode>tag: they support ssh, scp, rsync, etc
21:40<npmr>caker, i think i just need to check my bases with this one
21:40<npmr>caker, it's an image i haven't booted in a while
21:41<npmr>i think it may be doing stupid things like using nameservers that don't exist anymore
21:41<encode>i dont know if caker will mind if i write details of the company here
21:41<npmr>finally logged in
21:41<tag>Newsome: heh, gmailfs is cool shit.
21:41<caker> <-- it's there, if you want it
21:42<tag>caker: what's that?
21:42<caker>tag: a UML kernel
21:43<caker>encode: no, I think you've mentioned it before?
21:43<tag>no what's finnix
21:43<npmr>ha ha
21:43<npmr>changed nameservers, shutdown is proceeding normally
21:43<npmr>we'll see how the second boot goes
21:43<npmr>tag, finnix is fo0bar's distro
21:43<caker>npmr: ...... :)
21:44<tag>another debian distro
21:44<npmr>booting normally now
21:44<npmr>logged back in
21:44<caker>usually it'll hang up on sendmail/whatever, but not klogd..
21:45<npmr>caker, libpam-ldap
21:45<npmr>also, libnss-ldap
21:45<npmr>remote ldap server
21:45<caker>npmr: when it comes time to move the hosts to ldap, I'm coming to you there, buddy
21:45<npmr>too uml systems on the same host
21:45<npmr>two, even
21:45[~]mikegrb moves caker
21:46[~]caker directory accesses mikegrb
21:46[~]mikegrb likes
21:47<mikegrb>^ best boss ever
21:47<npmr>ha ha
21:47<mikegrb>I was pointing to myself, but you're not bad either caker
21:47[~]mikegrb runs
21:47|-|mode/#linode [+o caker] by ChanServ
21:48<npmr>go go gadget ChanServ!
21:48|-|mode/#linode [+oo mikegrb linbot] by caker
21:49<@linbot>npmr: *click*
21:49<encode>caker: yeah, i mentioned it before
21:49<encode>im still surprised how they manage to offer the storage so cheaply
21:53|-|FireSlash [] has joined #linode
22:10<npmr>Nov 29 22:56:26 egg kernel: Checking for /dev/anon on the host...Not available (open failed with errno 2)
22:10<npmr>caker, what's /dev/anon?
22:15<psykoyiko>it enables mmap for ubds
22:17<psykoyiko>..."the combination of /dev/anon and 'ubd=mmap' results in about a 25% decrease in host memory consumption. "
22:17<psykoyiko>^-- money.
22:18<FireSlash> <-- This is why I don't have a date :(
22:18<npmr>NOTE: At this point, 'ubd=mmap' is known to eat filesystems
22:18<psykoyiko>ha ha.
22:18<psykoyiko>but think of the memory you'll save!
22:20<npmr>i bet it eats babies, too
22:20<npmr>it could have been anyone!
22:25|-|jgoebel [] has joined #linode
22:26<jgoebel>mikegrb: what is Ticket 16265
22:29<@caker>npmr: ignore /dev/anon .. it's obsolete
22:30<npmr>and it eats babies
22:30<jgoebel>caker: hey
22:30<jgoebel>caker: can you look at ticket 16265 and tell me what it means
22:30<@mikegrb>jgoebel: spam from your Linode
22:31<@mikegrb>jgoebel: do you use sendmail.php to email webform replies?
22:31<jgoebel>mikegrb: sent where?
22:31<jgoebel>mikegrb: i have a php script we use for our contact forms
22:31<@mikegrb>the recipient was censored
22:31<jgoebel>mikegrb: it should only mail to our customers
22:31<@mikegrb>we received several reports
22:31<jgoebel>mikegrb: well, it would help to know if it's our customers or not
22:31<jgoebel>mikegrb: the robots actually fill out (parse, understand) the contact forms...
22:32<jgoebel>but it shuld only send the the address on file for that customer
22:32<jgoebel>ie, it shshould not be an open relay
22:32<@mikegrb>one was sent to a address the other to a address
22:33<jgoebel>mikegrb: can you get me full headers minux the e-mail addresses?
22:33<@mikegrb>you have them in the ticket
22:33<jgoebel>mikegrb: or is that what i should belooking at?
22:33<jgoebel>mikegrb: hmmm
22:34<jgoebel>i guess if it can forge the from header it can forge the to address as well
22:34<jgoebel>that sucks
22:34<@mikegrb>yeah, it's a pretty insecure script
22:34<jgoebel>how dumb is php's send function?
22:34<jgoebel>no, i wrote it
22:34<FireSlash>You can forge more or less everything with the current mail systems :/
22:34<@mikegrb>oh, so not sendmail.php
22:34<jgoebel>it set the addresses properly
22:34<jgoebel>mikegrb: no
22:34<jgoebel>mikegrb: hold one one sec, i'll pull up the sroce
22:35<@caker>jgoebel: perhaps greping your maillog for the IDs might provide more info on what happened on your end
22:35<jgoebel>caker: i think i know what's happening but i'm not sure how to fix it
22:35<@mikegrb>the headers do show it was from user www-data
22:36<@mikegrb>webserver logs from the time of the message could help though
22:37|-|jgoebel_ [] has joined #linode
22:37<jgoebel_>sorry disconnected
22:37<jgoebel_>it's contact forms on our websites i didn't know it was affecting more than our customers though (ie, the forms are set to only mail to ONE person)
22:37<jgoebel>it's contact forms on our websites
22:37<jgoebel>i didn't know it was affecting more than our customers though (ie, the forms are set to only mail to ONE person)
22:37|-|jgoebel [] has quit [Quit: jgoebel]
22:38<@mikegrb>jgoebel_: are there any legit or addresses?
22:38<jgoebel_>mikegrb: i don't believe so... though i'm not sure i know where people pop mail from
22:39<jgoebel_>mail($web_mail_to,"WEB: " . $subject,$msgbody,"From: $name <$email>");
22:39<jgoebel_>we provice $web_mail_to... ie, it's hardcoded
22:39<jgoebel_>the subject and msgbody comes from the form
22:39<jgoebel_>evidentally you can add arbitrary headers in the message body??
22:40<@mikegrb>From: " <better"
22:40<@mikegrb>Subject: [Fabulist] Link Submitted
22:40<@mikegrb>Subject: Huge market mover about to happen
22:40<@mikegrb>^ thats from the other report
22:40<@mikegrb>you could probably get the to address from your mail logs
22:40<@caker>What delineates headers from the message body? A blank line?
22:41<@mikegrb>Received: by (Postfix, from userid 33)
22:41<@mikegrb>id 50C6113EAF; Sun, 27 Nov 2005 04:51:42 -0500 (EST)
22:41<@mikegrb>caker: yeah
22:41<@mikegrb>caker: good idea
22:41<@caker>Could just throw a blank line in before the message body
22:41<jgoebel_>caker: yeah, that was my first thought
22:41<@mikegrb>see above
22:41<@mikegrb>excellent idea even
22:42<jgoebel_>caker: that would prevent the exploit?
22:42<jgoebel_>i'm reading the mail page on now
22:42<@caker>jgoebel_: it would prevent your MTA from interpreting any extra headers. They'd be treated as part of the message body
22:42<jgoebel_>mikegrb: yeah... someone posted about header injection on on the mail page :-)
22:42<jgoebel_>i'm reading what he did
22:42<jgoebel_>caker: yes, i need a multi-step fix
22:43<jgoebel_>caker: there are a lot of good tips on this page... he lots out agents not sneding HTTP_USER_AGENT and nonPOSTS
22:43<jgoebel_>ah, but i only handle posts already... so it's posting
22:43<@caker>and user agents can be changed trivially
22:43[~]mikegrb steps out for a cigarette
22:44[~]caker steps out for some candy
22:44<jgoebel_>caker: i know, but it's a good idea...
22:44<@caker>How would that stop anything?
22:44<jgoebel_>just one of many things he tried :-) i'm not proposing it as the solutoin
22:44<jgoebel_>i'm going to add a CR and some other things
22:48|-|jgoebel [] has joined #linode
22:48|-|jgoebel_ [] has quit [Quit: jgoebel_]
22:56<@mikegrb> <--
22:57<@caker>mmmm peanut brittle
22:58<psykoyiko>I think we all owe a big thank you to ssh tunnels.
22:58[~]psykoyiko shakes ssh tunnels' hand
22:58|-|jgoebel_ [] has joined #linode
22:58<jgoebel_>styupid php
22:58|-|jgoebel [] has quit [Quit: jgoebel]
22:58<@caker>jgoebel_: it's called screen+irssi
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
22:59<jgoebel_>caker: that sucks just about as bad
22:59<jgoebel_>caker: then i have to deal with lag all the time
22:59<@caker>lag for typing?
22:59<psykoyiko>does screen cause lag?
22:59<@caker>no lag here
22:59<npmr>none here either
22:59<jgoebel_>dialup causes lag
23:00<jgoebel_>this is very annoying though
23:01[~]linbot dispenses peanut brittle
23:02<jgoebel_>what log should be written to when a message is sent
23:03<npmr>your mta should log to /var/log/mail.log
23:03<jgoebel_>yeah, it's not doing that
23:03<npmr>most other components of a mail system don't log
23:03<npmr>are you using debian?
23:04<jgoebel_>yeah, i think there isn't enoguh free space on root
23:04<jgoebel_>darn spam
23:05<jgoebel_>ok the box is going down for a harddrive update
23:09<jgoebel_>now how long until the box powers down *waits*
23:11<FireSlash>Doing some security changes... just making sure: sit0, teql0 and gre0 are all "safe" connections and do not need firewalled, right?
23:12<jgoebel_>*copys his root filesystem*
23:16<jgoebel_>resizing desk
23:17<FireSlash>Hmmmmm... shutdown -r only shuts the system down.
23:19<jgoebel_>mikegrb: i'll update that ticket later tonight, i'm changing all the forms
23:19<@caker>jgoebel_: thanks for taking care of this
23:19<@mikegrb>jgoebel_: thanks a lot
23:19<jgoebel_>caker: well, it sucks as much for me as it does for you all
23:19<@mikegrb>most people don't bother doing anything until we've shutdown their linode
23:19<@mikegrb>nice to have someone hop right on it
23:20<@mikegrb>(or worse they fix it but never tells us and then complain when we've shutdown their linode)
23:20<jgoebel_>mikegrb: do you turn them off, or disable network or what?
23:20<@caker>or, we shut them down and then they say they've fixed it (with their node shutdown? .. um, no)
23:20<@mikegrb>jgoebel_: turn them off
23:21<@mikegrb>caker: yeah, that's a good one too
23:21<jgoebel_>i worked at a service center/extended warranty... and anyone who called with weird issues we'd ask if they're reloaded (from restore disks) their PC... no matter what, that was required to eliminate h/w as a possibility
23:22<jgoebel_>this guy called... i ask him and he said no... and i told him he'd have to do that... and he said "oh, well i did that"
23:22<jgoebel_>(he wasn't very happy before he lied)
23:22<jgoebel_>and i told him no way was i going to authorize anything today knowing he just lied to me... he'd have to at LEAST call back later and lie to someone else, but that he really neded to reload his box
23:22[~]mikegrb likes calling support when the cable modem is kaput
23:22<@mikegrb><them> ok, try rebooting your computer
23:23<@mikegrb>*mikegrb waits 30 seconds for his computer to "reboot"
23:23<@mikegrb><mikegrb> done
23:23<@mikegrb>it's the linux!
23:23<jgoebel_>yeah, don't tell them you don't run windows
23:23<jgoebel_>then it's always your fault
23:23<@mikegrb>my wife did this morning
23:24<@mikegrb>there reply was "well is there another computer that has windows? what's on the other computer?"
23:24<@mikegrb>"It's a mac."
23:24<@caker>I was all Mr SuperComputer when the comcast people came in to "install" cable-internet in my apartment
23:24<@caker>they were so lost
23:24[~]FireSlash just discovered nessus O.O
23:24<jgoebel_>*waits for his drive resize to complete*
23:24<@mikegrb>a long time ago when my parents got dsl, the phone guy comes to install it
23:24<jgoebel_>i just tell people I need a working ethernet jack
23:25<@mikegrb>house had cat5 jacks in all of the rooms but the kitchen and dining room
23:25<@caker>jgoebel_: exactly
23:25<FireSlash>Too bad it relies on a GUI :/
23:25<@mikegrb>run to patch panels in a closet
23:25<@mikegrb>he asked where the computer was that was getting the connection and I showed him
23:25<@mikegrb>he smiled
23:25<jgoebel_>though i'm totally moving to Mac now
23:26<jgoebel_>i've ordered three minis in the last week
23:26[~]caker wants a mactel laptop, possibly
23:26<@mikegrb>oh a mini
23:26[~]mikegrb will have to suggest that to wife
23:26<jgoebel_>mikegrb: beautiful, i'd recommend to anyone... but get 1g
23:26<jgoebel_>not 512mb
23:27<jgoebel_>osx is heavy but sweet
23:27[~]mikegrb has 512 in his ibook
23:27[~]mikegrb would prefer more
23:27<jgoebel_>i mean i'm living with 512mb, but ordering a 1g mini anytime
23:27<jgoebel_>it's usable, and too quiet to hear it swap, but i can tell it gets laggy
23:28<@caker>mikegrb: I wonder how many rooted-via-webmin Linodes we'll end up seeing
23:28<@mikegrb>or already have
23:28<@caker>webmin is probably as popular as phpbb +/-
23:29<@mikegrb>part of it's problem though is the insecurity it leaves in the daemons it configures
23:29<@caker>mikegrb: read the URL I pasted above?
23:29<@caker>webmin runs as root, too :(
23:30<@mikegrb>how long ago?
23:30<npmr>woe unto the ignoramus
23:33<@mikegrb>caker: what do you think of a scanner to look for linodes listening on the default webmin port, and generating a list, could then quickly manually double check with a browser that it was webmin listening on that port and open a ticket linking to the advisory
23:33<@mikegrb>"Fixed a possible remotely exploitable security hole caused by a bug in Perl." haha a bug in Perl
23:34|-|jgoebel [] has joined #linode
23:34|-|jgoebel_ [] has quit [Quit: jgoebel_]
23:36[~]FireSlash feels his geek factor has just gone up a bit more, now that he has used Lynx to run online port scanners ><
23:38<jgoebel>caker: you still here?
23:41<encode>how can i remove webmin from automatically starting on boot?
23:42<jgoebel>mikegrb: do a lot of people use that other insecure script?
23:43<@mikegrb>more then should ;)
23:43<jgoebel>seems lot of even the people writing their own scripts have been bit by this
23:44<jgoebel>everything thinks to protect the sender and recipient and that's it
23:44<@mikegrb>I think beside the other report we received this evening, only once have we had a report and it was within the last week or so
23:44<@mikegrb>don't know about the other one from this evening yet
23:44<jgoebel>mikegrb: you're talking about for me?
23:44<@mikegrb>no, there was another user hit by the same spammer we received a report about at the same time as yours
23:44<@linbot>... but there is the best shot of him or an exception for coinage or another version of the letter signed by somebody else or a sample in the forums labeled as for debian but it's the same or a checking for openssl... not found line before that or cream... which I haev not tried or LALUGS or something... they have a channel on freenode iirc or a law firm, a dentist, and something else or the vmstat command, mikegrb ...
23:44<@mikegrb>linbot: forget there
23:45<jgoebel>someone should write a post about this :-)
23:45<@mikegrb>can tell you that that particular spammer cropped up about august
23:45<jgoebel>or just link to the good stuff:
23:45<@mikegrb>seems to exclusively target form email scripts
23:45<jgoebel>mikegrb: light on fixes, but heavy on detailing the prob
23:45<encode>haha, i was reading what lin bot wrote, without realising that it came from linbot, tried to figure out what he was going on about
23:46<@mikegrb>and it's always pump and dump for vmtf
23:46<jgoebel>how do i lowercase something in php?
23:46[~]jgoebel uses Ruby all the time now
23:46<jgoebel>obj.downcase! :-)
23:47<jgoebel>strtolower, of course, why didn't i guess that
23:47<@mikegrb>his address lists are also always 98%
23:48<@mikegrb>that page looks good, haven't read it yet
23:48<jgoebel>mikegrb: what i read that i liked.... filtering body for individual headers... and if you allow the user to provide some header content (we send from the user, for example) make sure no
23:48<jgoebel>\r or \n are allowed
23:49<jgoebel>that way they can't add additional lines
23:55|-|jgoebel [] has quit [Remote host closed the connection]
23:55|-|jgoebel [] has joined #linode
---Logclosed Wed Nov 30 00:00:20 2005