Back to Home / #linode / 2006 / 07 / Prev Day | Next Day
#linode IRC Logs for 2006-07-07

---Logopened Fri Jul 07 00:00:42 2006
00:22<warewolf>I think I need to make my own internal root-zone for my home network, and chunks that hang off of it via vpn.
00:22<warewolf>I think I'll make it .x
00:22<warewolf>xabean.x, xasf.x, xell.x
00:22<warewolf>that'd be cool.
00:22<warewolf>now that I've figured out how to create "views" in bind
00:23<warewolf>it shouldn't be too difficult to keep it "internal" only
00:23|-|Eman [] has quit [Read error: Connection reset by peer]
00:24|-|Eman [] has joined #linode
00:39|-|internat [] has joined #linode
01:08|-|rh [~d5fe4967@] has joined #linode
01:09<rh>is there a way to have shutdown -r now to actually reboot the node instead of having to go to the control panel and click on the boot button?
01:11<warewolf>afaik, no
01:11<warewolf>it is best to reboot your linode via lish or the web interface
01:12<rh>why so?
01:12<warewolf>becuase it will actually boot it :)
01:12<rh>good answer indeed :)
01:12<warewolf>shutdown -r now will stop your linode, but not boot it.
01:13<warewolf>it's all in the way uml works, and linode's back end
01:13<warewolf>your linode is a 'process' running on the host
01:13<warewolf>when you shut it down, the process exits.
01:13<warewolf>how are they to know you want it back up?
01:14<Eman>cant it just exit with an errorlevel code or somesuch? even dos can do that
01:14<warewolf>you're thinking of -- isn't a kernel.
01:15<warewolf>in that case MSDOS.SYS and IO.SYS were the "kernel".
01:15<@caker>rh: there is a way -- one sec
01:15<warewolf>I stand corrected
01:15<warewolf>maybe it's as simple as 'schedule a boot' operation via the web console
01:15<@caker>well, it involves lish
01:16<rh>i was actually starting to brainstorm ideas on how to make it work
01:16<warewolf>I was right all along :)
01:16<warewolf>ssh 'reboot'
01:16<warewolf>*bamf* done
01:16<@caker>all the lish commands work via ssh
01:16<rh>duh :)
01:17<@caker>set up some host keys, and wallah
01:17<warewolf>I can see reasons why _not_ to have the linode back-end auto-boot a linode if it suddenly "exits".
01:17<warewolf>if your linode is crashing over and over again, it might get stuck in a loop
01:17<@caker>that and UML doesn't set a exitcode or some other way to check if the uml was shutdown vs rebooted
01:18<warewolf>I assumed as much
01:18<@caker>easy enough to make it touch a file or somesuch .. kinda what I was thinking once I get to the uml-stayup watchdog thing
01:18<warewolf>kinda difficult to set an exit code for a kernel; if it was implimented it'd be a bastardism.
01:18<rh>UML could just check if it has been rebooted N times in less than T seconds
01:18<warewolf>not UML, but linode's back end, rh.
01:18<warewolf>UML has no idea about it's previous boots
01:19<rh>but then i wonder how it'll stay powered off when you want it that way :)
01:19[~]caker performs his first zone import into the db
01:19<warewolf>it's like asking a light bulb "how many times have you been lit before?"
01:20[~]warewolf was about to use an automobile for that example, but then realised that today's cars have spiffy ECU (engine control units) that have non-volatile memory in them that can store info like that
01:21<@caker> <-- it works :)
01:22<rh>actually i cant see why automatic rebooting and poweroff shouldnt work
01:22<warewolf>oh poweroff works just fine
01:22<warewolf>coming back is the problem for UMLs
01:22<@caker>poweroff works .. rebooting I hard-coded UML to poweroff, because the host needs to clean stuff up, or configure a new devices/IPs, etc between reboots
01:22<@caker>stock UML can reboot itself just fine
01:23<warewolf>caker- it can?
01:23<rh>ah, ok.... that makes sense
01:23<warewolf>caker- the kernel process doesn't exit?
01:23<@caker>warewolf: hell yeah -- it just re-execs itself
01:24<warewolf>greagh crap
01:24<warewolf>this stupid cobalt raq needs 10.1 .. no wait it doesn't, that's only for the cobalt restore CD
01:24<@caker>actually, I think that's changed:
01:24<@caker>void reboot_skas(void)
01:24<@caker> block_signals();
01:24<@caker> UML_LONGJMP(&initial_jmpbuf, INIT_JMP_REBOOT);
01:24[~]warewolf raises an eyebrow
01:25<warewolf>that doesn't look like an exec :)
01:25<@caker>yeah, that's the bit that's different than I remember it
01:26<warewolf>er wrong window
01:26<rh>i wonder why they got rid of the exec
01:26<warewolf>what's that RFC1918 172.x.x.x segment?
01:26<warewolf>er 27
01:27<rh>to .31 ...
01:27<@caker>172.16/12, or -, according to google
01:27<warewolf>I'm just gonna use a /24 anyway :)
01:27<warewolf>for about half an hour heh
01:28<@caker>ok .. so left on the DNS Manager TODO: render out the zone, ability to run through named-checkzone, and then scripts for the nameserver nodes to suck everything in
01:29<warewolf>ok, dhcpd, nfs, and portmap all running
01:30<warewolf>*tries a nfs install of this cobalt box now*
01:34<linbot>New news from forums: [RFT] Zone AXFR Import Feature in DNS Manager (pre-beta) <>
01:36<warewolf>Red Hat install init version 9.0 using a serial console
01:36<warewolf>remember, cereal is an important part of a nutritionally balanced breakfast.
01:37<warewolf>ok so far so good, but why is the installer SEGVing
01:37<warewolf>I vaguely remember this happening once before
01:42<warewolf>no, it's something else
01:43<warewolf>eg, a bug
01:43<warewolf>I fixed it once before the last time I did this, but I don't remember what I had to do
01:43<rh>what are you installing, anyway?
01:43<warewolf>rh9 on a modified cobalt raq 4
01:44<warewolf>for work
01:44<warewolf>can't install anything later (atleast, at the moment)
01:48<rh>doesnt ring any bells here... we pretty much avoided red hat since 7.3 in favour of the as/es lines...
01:54<warewolf>hmm, I think I need to use a different kernel
02:08<warewolf>yeah, the in-prom kernel is not going to work correctly
02:11[~]warewolf hopes symantec didn't mess with the prom kernel or hardware in general on this box too much
02:17[~]warewolf dumps orig rom image and flashes rom
02:22<warewolf>Yep! That was the problem.
02:22<warewolf>new rom works like a champ.
02:25<warewolf>oh, that's right I had to use TWO ethernet interfaces to install
02:25<warewolf>man, is this install ever tricky
02:30<rh>you should document it somewhere for next time :)
02:34<warewolf>I am, step by step :)
02:41|-|womble [~mpalmer@] has quit [Quit: What's behind the round window...]
03:00|-|netjackal [~netjackal@] has quit [Ping timeout: 480 seconds]
03:17<linbot>New news from forums: open source Enterprise backup system? in /dev/random <>
03:19|-|JavaWoman [] has quit [Quit: Logging out]
03:25|-|JavaWoman [] has joined #linode
03:55|-|rh [~d5fe4967@] has quit [Quit: CGI:IRC (Ping timeout)]
04:06|-|Battousai [] has quit [Ping timeout: 480 seconds]
05:55|-|FireSlash [~FireSlash@] has joined #linode
06:08|-|FireSlash [~FireSlash@] has quit [Read error: Connection reset by peer]
06:35|-|Dreamr_3 [] has joined #linode
06:40|-|Dreameer3 [] has quit [Ping timeout: 480 seconds]
06:45|-|Spads [] has joined #linode
07:08|-|internat [] has quit [Quit: This computer has gone to sleep]
07:24|-|KE3VIN [] has joined #linode
07:26<KE3VIN>Urm, we seem to be having latency issues with HE .. again..
07:29|-|KE3VIN [] has quit [Quit: KE3VIN]
07:32|-|konoko [] has quit [Ping timeout: 480 seconds]
07:36|-|linville [] has joined #linode
08:04|-|konoko [] has joined #linode
08:08|-|jekil [~alessandr@] has joined #linode
08:09|-|bakins [] has joined #linode
08:10<bakins>cacker: you awake?
08:10<bakins>caker, I mean
08:16<anderiv>haven't seen him yet today.
08:18<linbot>New news from forums: Anoher cancellation, no refund in Sales Questions and Answers <>
08:19<bakins>the cancellation question is me. Seems to be a running theme...
08:19<Spads>it's still early morning in the states
08:20<Spads>give them time to wake up
08:20<bakins>it's 9:20 here...
08:20<Spads>they stay up pretty late most nights
08:20[~]anderiv was at work before 6:30 today :-(
08:20<@caker>bakins: I've got no problem giving you a refund, but please state the facts:
08:20<bakins>caker, I just clarified in the forum...
08:20<@caker>bakins: you didn't cancel until after you were already billed
08:21<@caker>PAYMENTDATESTAMP: 2006-06-01 06:15:33
08:21<@caker>Cancel Date: 2006-06-01 11:52:08
08:21<bakins>caker: not as I saw it, unless you charge at midnight and I didn't get notification until after I cancelled.
08:22<@caker>bakins: also, a dispute costs us an additional $75 bucks -- so I'd appreciate you giving us a chance to credit you
08:22<anderiv>caker: ooh 75 bones...that sucks.
08:22<bakins>caker: sure, need it soon.
08:24<bakins>caker: dispute will be filled today.
08:24<bakins>have already started the process, however, I can stop it, if I get a refund.
08:25<@caker>bakins: one sec
08:27<bakins>cool. Thanks.
08:28|-|bakins [] has quit [Quit: CGI:IRC]
08:44|-|internat [] has joined #linode
09:07|-|emcnabb [] has joined #linode
09:40|-|Spads [] has quit [Ping timeout: 480 seconds]
09:46|-|Spads [] has joined #linode
09:58|-|fapestngd [~james@] has joined #linode
09:58|-|fapestngd changed nick to fapestniegd
09:59<fapestniegd>HE issues again?
10:00<emcnabb>looks like that to me
10:01<@caker>"A down cross-country link is causing some traffic shifts at this time. The line provider has been contacted and is working on the issue. I do not have a firm ETA from them as of yet for resolution." <-- HE
10:01|-|Spads [] has quit [Ping timeout: 480 seconds]
10:02<fapestniegd>caker: thanks, the congestion isn't as bad as the other day, it's just causing lag. services are still available.
10:07<emcnabb>what ended up happening earlier this week?
10:12|-|Rifkin [~rifkin@] has joined #linode
10:14<neale>well, on Monday caker balanced a bowling pin on his nose for a full 10 minutes
10:14<emcnabb>awesome! :-)
10:14<neale>yeah, you really missed out
10:14<neale>it was excitement
10:15<emcnabb>I'll be more specific this time: how about with the network issues on the 4th?
10:17[~]neale shrugs
10:21<Rifkin>Hello, I have a newbie question about my linode
10:21<Rifkin>anyone awake?
10:25<fapestniegd>Rifkin: what up?
10:26<Rifkin>My load average
10:26<Rifkin>is it just my linode, pr all nodes on the host I'm sharing?
10:27<fapestniegd>As a rule, just yours.
10:27<Rifkin>that's strange
10:27<fapestniegd>But load could mean IO as well, which could be inpacted by other hosts.
10:27<Rifkin>top shows a load average consistantly around 0.6, but no processes using any CPU
10:28<fapestniegd>In my experience, this is negligible
10:28<fapestniegd>load isn't just cpu
10:28<Rifkin>so it's normal to have a load average like that, even when nothing is running?
10:28<fapestniegd>load is the amount of time the average process spends "waiting for something"
10:29<fapestniegd>the key concept here is the "something"
10:30<fapestniegd>0.6 seems a little high for a completely idle box, but not unreasonably so.
10:31<Rifkin>So it's not anything I'm doing? Is there any way I can find out what's causing it?
10:31<fapestniegd>Without knowing more about your setup it would be impossible to tell.
10:31<fapestniegd>Is it effecting performance?
10:31<Rifkin>I don't think so
10:31<fapestniegd>You really shouldn't focus on numbers like load average as a performance measurement
10:32<fapestniegd>unless services are slowing down and you log in and see a load of 23.00
10:32|-|kokoko [] has joined #linode
10:32<Rifkin>what's a better number to look at?
10:33<neale>cpu usage
10:33<fapestniegd>vmstat/iostat/sar output
10:33<neale>yeah or vmstat
10:33<fapestniegd>they will tell you what your system is actually waiting on, which is causing the load
10:34<Rifkin>hmm I don't have iostat or sar installed
10:34<fapestniegd>what OS?
10:34<fapestniegd>apt-get install sysstat
10:35<fapestniegd>will give you sar,iostat, and mpstat
10:35<neale>vmstat may be all you need though
10:36<neale>back in the day that's all we had
10:36<neale>and we liked it
10:36<Rifkin>yeah vmstat shows 99% idle cpu. I was confused at the discrepancy between that and load average
10:36<neale>run "vmstat 2" and watch the numbers change
10:37<neale>you may have something swapping
10:37<fapestniegd>load can be caused by i/o (bi/bo) or context switching (cs) as well
10:37<Rifkin>nope, nothing. 100% idle. 0 swap
10:37<neale>and no io?
10:38<Rifkin>I show maybe 1 block out every 10 sec or so
10:38<neale>well, so that's great, your machine should be nice and responsive :)
10:38<Rifkin>Yeah, I haven't noticed any slowness, I was just confused where this high load average was coming from
10:39<neale>I've noticed the load average on my linode is typically around 0.6
10:40<neale>so, just don't worry about it until it gets to 2
10:40|-|konoko [] has quit [Ping timeout: 480 seconds]
10:40<Rifkin>ok I'll just ignore it
10:41<npmr>0.6 is not a high loadavg
10:42<Rifkin>I thought a completely idle linode would have a zero load average
10:42<neale>no, but it's probably higher than you'd see on a dedicated machine with the same processes.
10:42<neale>Rifkin: it would
10:43<neale>but there'd be no way to see it at 0
10:43<neale>sort of like quantum mechanics
10:43<neale>the act of observing it alters its state
10:43<npmr>which processes have the highest cumulative cpu time?
10:43<anderiv>neale: well put :-)
10:43<Rifkin>when my home linux box is idle and I look at the load av, it shows zero
10:43<neale>Rifkin: it's rounding ;)
10:44<Rifkin>ok, but it's nowhere near .6
10:44<npmr>Rifkin, that doesn't make 0.6 high
10:44<neale>I think the answer is "expect an unloaded linode to have a little bit higher load average"
10:45<neale>I spent a good deal of time trying to get mine down at first
10:45<neale>thinking something was wrong
10:46<fapestniegd>keep in mind that one state 'D' process will bring you load to 1.00, because that process is in an uniteruptible I/O wait state. It's not actually doing anything, it's just waiting. So load is decieving.
10:46<npmr>min loadavg on mine over the last 12 months is 0.47, as sampled by munin
10:46<fapestniegd>it's just not a good metric
10:46<npmr>well, the instantaneous load, yes
10:46<npmr>the loadavg is time-weighted average of instantaneous data
10:47<fapestniegd>you mean a rolling average of instantaneous data
10:47<neale>the first thing you learn in sysadmin 101 is that load average isn't what you think it is :)
10:47<npmr>that was a piece of math i never studied, so i don't know the precise terminology
10:48<neale>typically the load average is only an indicator that things are getting backed up
10:48<neale>so if you see it spike, then you need to look around to see what's up
10:48<neale>otherwise, I typically ignore it
10:49<neale>and in my experience, spike in load average usually means lots of email
10:49<Rifkin>I don't have a problem ignoring the load average, it would just be interesting to know what's causing it
10:49<fapestniegd>If I had to guess, you've got a service that is polling something causing it.
10:50<neale>that'd do it
10:50<fapestniegd>shut down your applications, one at a time and see what happens to the load.
10:50<neale>okay so here's a linode I have
10:50<fapestniegd>then bring them back up and see if it reverses.
10:51<neale>running sshd, syslogd, klogd, exim4, inetd, vsftpd, cron, mathopd, and getty
10:51<neale>all of those are blocking on an accept or select
10:51<neale>and the load average is in fact 0
10:51<neale>top - 11:51:22 up 186 days, 8:37, 1 user, load average: 0.00, 0.00, 0.00
10:51<Rifkin>I'm just running lighttpd, ruby/rails, a couple cron jobs. Everything is showing sleep status except top and sshd
10:51<neale>so I would side with fapestniegd on this one: you've probably got something polling something
10:52<neale>right but the S only means right at that instant
10:52<neale>it might still be running
10:52<neale>try stracing those
10:52<Rifkin>strace which?
10:52<neale>your server processes
10:53<neale>top will affect the load average too, BTW
10:53<fapestniegd>strace -p <PID>
10:53<Rifkin>ok, got to install strace, which package is it in?
10:54<fapestniegd>apt-get install strace
10:55<neale>"apt-cache search programname" will typically show you what package something is in.
10:55<Rifkin>ok, the cgi dispatcher is just waiting in select()
10:55<Rifkin>it's not doing anything
10:55<neale>what's the last argument in the select call?
10:56<neale>okay so it's blocking forever
10:56<neale>try another one
10:56<Rifkin>select(5, [0 4], [], [], NULL <unfinished ...>
10:56<neale>the last one is a timeout
10:56<neale>some things will say "wait 100ms for something to happen"
10:56<Rifkin>the main ruby process is doing the same
10:56<neale>and then go check a bunch of stuff
10:56<Rifkin>I only have one dispatcher thread running
10:56<neale>check lighttpd
10:57<Rifkin>a ha
10:57<Rifkin>it's polling
10:57<neale>BTW, I could not get lighttpd to stop crashing on my box
10:57<neale>every once in a while it would just lock up
10:57<neale>waiting for a write to complete
10:57<Rifkin>it's working pretty well for me
10:57<neale>yeah, when it worked it was great
10:57<Rifkin>but strace shows one of these every second:
10:57<neale>and I was using fcgi
10:57<Rifkin>poll([{fd=4, events=POLLIN}, {fd=-1}, {fd=-1}, {fd=-1}, {fd=-1}, {fd=-1}], 6, 1000) = 0
10:58<neale>okay, so there's the timeout
10:58<neale>poll is actually pretty much like select
10:58<Rifkin>so lighttpd is being bad
10:58<neale>but that last value means it's doing something every once in a while, which will raise your load average
10:58<neale>no, that's not bad
10:58<neale>it's just doing stuff :)
10:58<neale>when you thought it wasn't.
10:59<Rifkin>there's no reason for it to poll like that though, it should just wait on a select, shouldn't it?
11:00<neale>use the source, luke
11:00<neale>maybe it's doing garbage collection
11:00<neale>or updating a log file
11:00<neale>who knows
11:00<Rifkin>I'm not going to comb through the lighttpd source :/
11:00<neale>neither am I :)
11:01<Rifkin>but I will look through the lighttpd forums to see if anyone else is complainig about this
11:01<Rifkin>thanks for helping me find it
11:01<neale>dude, it's not a bug
11:01<fapestniegd>are you using any AJAX?
11:01<neale>it's just that lighttpd is doing something periodically
11:01<Rifkin>no, no ajax calls, yes
11:01<neale>that's pretty normal
11:01<Rifkin>no ajax calls yet
11:02<fapestniegd>it seems using poll() is in the design of lighthttpd
11:02<fapestniegd>it's a "feature"
11:02<warewolf>and don't forget
11:02<neale>you need to understand what poll() does
11:03<neale>poll() is like select() 2000
11:03<warewolf>http 1.1 complaince is a "feature" too!
11:03<warewolf>and CGI, and GET and POST ....
11:03<neale>man poll
11:04<neale>lighttpd probably uses poll because if it used select it could only have a maximum of 255 simultaneous connections
11:04|-|kokoko [] has quit [Ping timeout: 480 seconds]
11:04<fapestniegd>so we've determined here that it's lighttpd, and that it is doing exaclty what it was written to do, causing a mildly-low load of 0.6
11:05<Rifkin>my little server will never have that many connectins. I wonder if I can tell lighty to use select instead
11:05<neale>Rifkin: the use of poll() is not what's causing the load average to be non-zero
11:06<neale>Rifkin: it's the timeout value
11:06<neale>Rifkin: if you modified the code to not use a timeout in poll, and killed every other process on your machine, I'm sure you'd see your load average go back down.
11:07<neale>I'm growing weary of your witch hunt, though.
11:07<Rifkin>or just increase the timeout to something more than 1 sec
11:07<neale>this reminds me of the first time I saw a Unix machine
11:07<neale>it was just sitting there, nobody was doing anything on it
11:08<fapestniegd>neale: i agree it seems we're tuning the idle loop here
11:08|-|konoko [~Vampire@] has joined #linode
11:08<neale>and the hard drive light was flashing periodically
11:08<neale>and I was all "why is it accessing the hard drive? nothing's happening!"
11:08<neale>it really bothered me
11:08<neale>until I learned that Unix can do more than one thing at once, and that it was just tidying things up since the user wasn't asking it to do anything else.
11:09<neale>So you can think of lighttpd doing the same thing. It's just keeping stuff neat and clean since it's not busy with anything else.
11:09<fapestniegd>yeah, the irony being that the light was flashing *because* the system was idle
11:09<Rifkin>I know
11:12|-|sneakums [] has quit [Read error: Connection reset by peer]
11:12|-|sneakums [] has joined #linode
11:13|-|fapestniegd [~james@] has quit [Quit: need sleep... ]
11:20|-|spr [~spr@] has joined #linode
11:29|-|Rifkin [~rifkin@] has quit [Quit: leaving]
11:44|-|unknown [] has quit [Quit: ..(cyp): [BX] Who ate my nuggets?!]
11:44|-|adamg [~misthos@] has joined #linode
12:03|-|kokoko [] has joined #linode
12:09|-|konoko [~Vampire@] has quit [Ping timeout: 480 seconds]
12:09<neale>does anybody actually use CNAMEs anymore?
12:09<anderiv>neale: yes
12:10<bendy24>handy for redudancy
12:10<neale>I never understood the point. It seems like a hack for people who are too lazy to write macros.
12:10<warewolf>yes, but don't use cnames in anything related to mail :)
12:10<anderiv>warewolf: hehe - I found that one out the hard way many years ago :-)
12:10<warewolf>me too
12:10<warewolf>couple months ago
12:11<anderiv>MX records have to map to a valid A record.
12:11<warewolf>coworker set the cname and I was bashing my head against the wall "WHY ISN'T IT WORKING WHY IS SENDMAIL ACTING WRONG WHAT DID HE DO TO SENDMAIL"
12:11<warewolf>anderiv: right :)
12:11<anderiv>that's about the extent of my DNS knowledge, though :-)
12:11<neale>you dudes all use BIND?
12:11<warewolf>I do
12:11<anderiv>neale: I use whatever uses...
12:12<neale>oh hey, didn't caker set up some sort of linode secondary nameserver dealie?
12:12<npmr>it's in beta
12:12<neale>I'd totally beta test that one
12:12<npmr>secondary, primary, or both
12:12<warewolf>just primary folks
12:12[~]bendy24 cnames warewolf
12:12<warewolf>no slaves yet
12:28|-|afv-13 [] has joined #linode
12:31<warewolf>sweet, I can go home now.
12:31<warewolf>another uh .. 13 hour day.
12:31<afv-13>join the club
12:32<warewolf>lemme add up my hours
12:32<warewolf>63h this week
12:33<bendy24>thats what sendmail does to you
12:34<neale>warewolf: you taking off for the day?
12:40<neale>guess so :)
12:57<emcnabb>caker, mikegrb: my linode randomly powers off (a least once a week, if not more)
12:57<emcnabb>there's nothing I can see in the logs that helps narrow down the problem
12:58<emcnabb>what should I do next to try to figure out what's going on?
13:02<emcnabb>or if anyone else has an idea, I'd love to hear it :-)
13:04<TheFirst>checked the logview i assume?
13:06<emcnabb>I actually haven't
13:07<emcnabb>where can I access it?
13:07<anderiv>emcnabb: lish
13:08<emcnabb>thanks, I'll check it out
13:08<bendy24>i had that problem with my xennode
13:12|-|Battousai [] has joined #linode
13:21<neale>if [ $$ -gt 12000 ]; then halt
13:22<emcnabb>This is what I see when running ''ssh logview"
13:22<neale>emcnabb: you don't have anything like that running from cron, do you?
13:22<emcnabb>it's a kernel panic
13:22<emcnabb>I wonder if I need to be running a different kernel
13:22<bendy24>cobra login: Kernel panic - not syncing: Segfault with no mm
13:22<bendy24>yeah, try .17
13:24|-|Netsplit <-> quits: neale, Eman
13:25<emcnabb>thanks; I'm going to try it right now
13:26|-|Netsplit over, joins: Eman
13:26|-|Netsplit over, joins: neale
13:38|-|cout [] has quit [Ping timeout: 480 seconds]
13:46|-|cout [] has joined #linode
13:51<neale>my linode is now being rated
13:51<neale>seeing as how I wrote a whole new client
13:52<Battousai>in fortran?
13:53<neale>oh man I totally should.
13:53<neale>I wonder if UCB LOGO has network routines
13:58<neale>I could call out to netcat from logo...
14:04<neale>I guess I maybe found an parser problem in caker's zone-xfer-testbench?
14:05<neale>zone transfer, ns, check record 6
14:11<@mikegrb>neale: isn't that the a record for
14:11<@mikegrb>you silly kanigit
14:11<neale>well I don't know what it's for, there's no hostname!
14:11<@mikegrb>see: 3600 IN A
14:12<neale>I see this:
14:12<neale> HOST
14:12<neale> TARGET
14:12<neale> TYPE A
14:12<@mikegrb>that's the a record for
14:12<neale>maybe that's expected?
14:12<@mikegrb>if you look in your zonefile it will be similiar
14:13<neale>well it seemed unusual since everybody else had a FQDN for the HOST field
14:13<@mikegrb>6th line down from the soa line at the top where it has the raw stuffs
14:13<neale>my zonefile has a FQDN
14:14<@mikegrb>but bind won't give it
14:14<neale>I mean, whatever
14:14<@mikegrb>that's the raw output from bind's afxr command
14:14<neale>we don't need to argue about this, I just thought it might be a problem
14:14<@mikegrb>yeah, it's normal and expected
14:14<neale>I'm running NSD instead of BIND so figured maybe it was triggering some funkiness
14:15<@mikegrb>with bind you would just leave it off in the zonefile
14:15<@mikegrb>but I believe you can add it and it will just cut it off internally
14:16<bendy24>im #2!
14:18<neale>so I actually wrote a big honkin' m4 macro thing to generate zone files
14:18<neale>the input to it looks like:
14:18<neale>ZONE(, FOZZIE)
14:18<neale>HOST(, FOZZIE)
14:19<neale>HOST(, FOZZIE)
14:19<neale>HOST(, HOME)
14:19<neale>HOST(, HOME)
14:19<neale>HOST(, HOME)
14:19<warewolf>m4 is the shit
14:19<neale>HOST(, FOZZIE)
14:20<neale>and it spits out a zone file
14:20<neale>I only mention this because it seems like the sort of thing somebody might be interested in.
14:21<neale>yeah I'm pretty happy with it
14:22<neale>hey warewolf, I wrote a new rml client in python
14:22<warewolf>the whitespace language
14:22<neale>it's sitting in my mail queue waiting for 31 minutes and 18 seconds
14:22<neale>which by the way seems like an excessively long whitelist period
14:23<warewolf>my greylist is 1h
14:23<warewolf>so it'll be another 28 minutes
14:23<warewolf>nothing against you, I get a lot of spam.
14:23<neale>s/wrote/wrote you/
14:23<neale>well I didn't have some perl library your thing needed
14:24<warewolf>if you don't have that, you must be a debian user, without half the stuff that is supposed to come with perl
14:24<@mikegrb>ha ha
14:24<warewolf>redhat does it too
14:24<warewolf>redhat does itI hate it
14:25<neale>does the 1h timeout make a measurable difference?
14:25<warewolf>neale: for me it does
14:25<neale>or did it just seem like a good idea?
14:25<warewolf>anything shorter, and spam from the same "from" address, but not necessarially the same message, got through.
14:25<neale>yeah, whatever it's called, I quit doing perl at 4.01
14:25<warewolf>like I said, I get a *lot* of spam :)
14:26<warewolf>that's like, smithsonian historic
14:26<@mikegrb>omg, not everybody uses perl
14:26<warewolf>omg, mikegrb hates on warewolf
14:27<warewolf>and it's actually getting old.
14:27<warewolf>I don't know what I did to make you so hostile towards me, but I'd like it to stop.
14:28<@mikegrb>so because I say not everybody uses perl I'm being hostile towards you?
14:28<@mikegrb>I'm sorry but I didn't realize facts offended you so much
14:28<warewolf>it was the sarcasm.
14:28<@mikegrb>I will reframe from stating facts in the future lest it offend you.
14:28<warewolf>again, sarcasm.
14:28<@mikegrb>I was being serious.
14:30<neale>okay boys, that's enough.
14:31<warewolf>neale: hahah I see your "testing" your client
14:31<neale>sheesh if I'd know it'd result in this I wouldn't have bothered writing the client ;)
14:31<warewolf>neale: unrelated.
14:32<neale>sweet! I'm coming in above nova1!
14:32<warewolf>are you supplying fictional data?
14:32<neale>yep, that was me.
14:33<warewolf>dun do dat pleeeese
14:33<warewolf>or atleast, when you're done, delete that host
14:34<neale>I mean, I don't think so.
14:34<neale>what looks wrong?
14:34[~]warewolf rewinds
14:34<neale>I did
14:35<warewolf>I saw you toss a few results back at RML, which it told me it didn't expect (you tried to submit data for a host that didn't exist)
14:35<warewolf>those were fine
14:35<warewolf>didn't do anything
14:36<neale>well if something looks wrong I'm happy to look into it
14:36<neale>but AFAIK I'm sending you stuff verbatim out of proc
14:36<warewolf>what I /was/ concerned about, was your comment "whee I'm coming in above nova1" (paraphrased) -- if you were supplying fictional data that showed an excess io rate, I'd prefer you didn't
14:36<warewolf>that's all
14:36<warewolf>no biggie
14:36<warewolf>beware one thing
14:36<neale>yeah I had host and host_system switched
14:36<warewolf>there is an integer roll over in /proc/io_status
14:37<warewolf>(yeah, I shoulda named those 'linode' and 'host')
14:37<warewolf>actually I might take care of the roll-over on the back end, lemme check
14:37<neale>that was more a joke because nova1 has ginormous numbers
14:38<warewolf>that's caker's box
14:38<neale>it's caker's box
14:38<warewolf>it does big things for linode, or something.
14:38<neale>I guess he does a lot of junk on it
14:38<Eman>lots of porn torrents take a lot of io :p
14:40<warewolf>neale: just to be clear, I have -no- problem with you writing your own (compatible) client.
14:41<warewolf>infact I'll probally make it available on the website (with credit to you) once it gets past my greylisting.
14:57[~]warewolf -c sleep &
15:20<neale>I'm going to set up a 48 hour delay on my greylisting.
15:21<neale>I bet that'll get rid of all my spam!
15:23<@mikegrb>better to be safe, go with 72
15:46<neale>sage advice
15:47|-|FireSlash [] has joined #linode
15:52|-|FireSlash [] has quit [Quit: ]
16:00|-|afv-13 [] has quit [Quit: leaving]
16:06|-|linville [] has quit [Quit: Leaving]
16:11|-|emcnabb [] has quit [Quit: Leaving]
16:22|-|neale [] has quit [Write error: connection closed]
16:45|-|adamg [~misthos@] has quit [Ping timeout: 480 seconds]
17:50|-|jekil [~alessandr@] has quit [Ping timeout: 480 seconds]
18:13|-|spr [~spr@] has quit [Quit: "Ah, the bounce has gone from his bungie"]
18:27<linbot>New news from forums: Fedora Core 4 Support? in Feature Request/Bug Report <>
18:30|-|miallen [~458ec4aa@] has joined #linode
18:32<miallen>How do I set my linodes hostname?
18:32<miallen>I'm using debian small.
18:33<miallen>Will putting the hostname in /etc/hostname have any effect?
18:34<Redgore>miallen: after a reboot yes
18:34|-|ElectricElf [] has quit [Remote host closed the connection]
18:34|-|ElectricElf [] has joined #linode
18:36<@mikegrb> '
18:36|-|Dreameer3 [] has joined #linode
18:37<miallen>Ok, here goes ....
18:41|-|ElectricElf [] has quit [Quit: ]
18:42|-|Dreamr_3 [] has quit [Ping timeout: 480 seconds]
18:42|-|ElectricElf [] has joined #linode
18:43|-|ElectricElf [] has quit [Quit: ]
18:43|-|ElectricElf [] has joined #linode
18:43|-|ElectricElf [] has quit [Quit: ]
18:43|-|ElectricElf [] has joined #linode
18:44|-|AmarO [~AmarO@] has joined #linode
18:45<AmarO>i have a question
18:45<miallen>Thanks Redgore.
18:46<Redgore>AmarO: well ask itr
18:47<AmarO>is allowed the use of manifold server IRC in the VPS?
18:47<Redgore>as long as you dont brweach copyright your fine
18:48<AmarO>very good
18:49<AmarO>Linode 100 is available for the purchase at this moment?
18:50<eFUDd>linbot, linode avail?
18:50<linbot>mikegrb: Linode100 - 0, Linode150 - 4, Linode200 - 5, Linode300 - 2, Linode400 - 2
18:50<eFUDd>welll, there ya go then.
18:50[~]eFUDd should reboot sometime to get that memory
18:51<@mikegrb>there eFUDd goes again, trying to be helpful
18:51[~]eFUDd fails
18:51<eFUDd>hey, i got a raise+bonus+grade bump+promotion today!
18:51<@caker>mikegrb: where have all the cowboys gone?
18:51<@caker>eFUDd: grats
18:51[~]eFUDd nods
18:52<@mikegrb>caker: Where is my happy ending?
18:53<AmarO>then, they could reserve when there is availability?
18:56<AmarO>or buy, and ... wait
18:59<AmarO>i have another questions
19:00<AmarO>how many IPs I can tenér in my plan VPS?
19:00<AmarO>in Linode 100
19:00<linbot>AmarO: Linode100 - 0, Linode150 - 4, Linode200 - 5, Linode300 - 2, Linode400 - 2
19:01|-|AmarO [~AmarO@] has quit [Quit: AmarO]
19:03|-|netjackal [~netjackal@] has joined #linode
19:03|-|patricio [LinodeJava@] has joined #linode
19:03<patricio>aca estoy
19:04|-|AmaRo [~AmaRo@] has joined #linode
19:05<eFUDd>co.. oh, nevermind.
19:06<eFUDd>(Please do not feed the efudd troll.)
19:06<AmaRo>how many ips i can have in linode 100 ?
19:07<@mikegrb>2 unless you have a valid reason for more
19:08<patricio>for more ircs
19:08<AmaRo>for example, for server IRC?
19:09<patricio>one ip... one irc..
19:10<AmaRo>all right, but ... if I want to have more servers IRC, i need more IPS
19:10<AmaRo>its good reason for that?
19:12<AmaRo>mikegrb: if i want more IPS for servers IRC, its a good reason for that?
19:12<AmaRo>or not?
19:13<patricio>we pay the excedent ips
19:13<@mikegrb>because ARIN has strict rules for IP allocation and we have to go by their rules or else all of our IPs may be taken away
19:14<patricio>who is ARIN?
19:15<@mikegrb>the organization responsible for IP delegation in north america
19:15<patricio>ah okey
19:16<patricio>we have a vps in
19:16<miallen>Has anyone tried GoDaddy's Turbo SSL certs for HTTPS?
19:16<patricio>I need 5 ips
19:16<miallen>They only contain domain information. No business address or location. Could that be a problem for some browsers?
19:17<@mikegrb>miallen: as long as their CA cert is truested, it shouldn't be
19:18<miallen>They're only $20/yr which is so much cheaper than everyone else I'm worried they must be defective.
19:19<miallen>mikegrb: Well no doubt that must be the case or they wouldn't be worth $2 :->
19:21<@mikegrb>er trusted
19:41|-|AmaRo [~AmaRo@] has quit [Quit: AmaRo]
19:42|-|patricio [LinodeJava@] has quit [Quit: patricio]
19:49|-|Kurt [] has joined #linode
20:15|-|thoth39 [~hm@] has joined #linode
20:28|-|dscott [~dscott@] has joined #linode
20:36|-|lightern [] has left #linode [Leaving]
20:40|-|dscott [~dscott@] has quit [Quit: dscott]
20:42|-|Kurt2 [] has joined #linode
20:50|-|Kurt [] has quit [Ping timeout: 480 seconds]
20:55|-|Kurt2 [] has quit [Ping timeout: 480 seconds]
21:58<eFUDd> <- is that ferreal?
22:10|-|netjackal [~netjackal@] has quit [Ping timeout: 480 seconds]
22:25[~]warewolf sighs
22:26<warewolf>I'm just a wee bit pissed at the folks at my office -- not the people I work with directly, they're awesome
22:26<warewolf>but the people who decided that contractors like myself are scum, and nolonger deserve to park at my office, and the employees who were relocated get priority.
22:27<thoth39>There's some logic there.
22:29<warewolf>it's their fault for not having a proper disaster recovery plan.
22:29<warewolf>and some people (contractors) are now out of a job.
22:29<warewolf>thankfully I get to keep mine, atleast for the forseeable near fututre
22:29<warewolf>but it's still messed up
22:36<linbot>warewolf: Your current monologue is at least 9 lines long.
22:37[~]warewolf -c sleep & # again, after getting woke up by the office
22:55|-|brocktice [] has joined #linode
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
23:06<brocktice>I'm really wishing for more storage... and my search of the Forums has come up rather empty... I would love to throw some money at Linode for disk space but the cost is currently too prohibitive. Is there no way to hook up fiber channel storage arrays to the nodes for some affordable storage?
23:18<TheFirst>i'm in the same wishing boat
23:19<TheFirst>caker said a while back he had plans to increase storage at some point....dunno if that's still the case or what's involved (more hds or something) but...
23:20<npmr>there is a rather large plan
23:21<npmr>involving a lot more than just storage
23:22<brocktice>npmr... I think I already have the second largest plan
23:22<brocktice>and I need (read: would like) orders of magnitude more storage
23:23<brocktice>I do research where we generate tons of data, and use all kinds of storage solutions... it's just not that difficult or expensive anymore
23:29<brocktice>Maybe I should look into putting some fileservers in ThePlanet
23:29<brocktice>selling space to Linode users :)
23:41<brocktice>I wonder if there's a market?
23:42<brocktice>I *am* intrigued by macminicolo. It's cute if not necessarily compelling
23:58<linbot>New news from forums: 'Storage' Linodes? in Feature Request/Bug Report <>
---Logclosed Sat Jul 08 00:00:30 2006