Back to Home / #linode / 2006 / 06 / Prev Day | Next Day
#linode IRC Logs for 2006-06-11

---Logopened Sun Jun 11 00:00:17 2006
00:02|-|Spads [] has joined #linode
00:17|-|flatronf700B [~flatronf7@] has joined #linode
00:21<warewolf>I wonder if redhat killed my RHEL subscription (it is about to, or has expired) yet.
00:22<TheFirst>RHEL...what a joke
00:23<warewolf>... right.
00:23<warewolf>would you like some debian eliteism, or gentoo -O1^1024 speed-enhancing racing stripes with your fries sir?
00:24<TheFirst>heh no eliteism about it...last job ran rhel and their support and updates were a joke
00:24<TheFirst>support was 'install this to fix your problem and we cant support you anymore' and their updates consistently broke crap
00:24<warewolf>something must have been set up wrong then, because I've had only one problem, and it was my own dman fault
00:25<TheFirst>the big problem was known bugs in samba
00:25<TheFirst>that had long since been fixed
00:25<TheFirst>and they had rpms for
00:25<warewolf>um right. You have to use RPMS from Red Hat.
00:25<TheFirst>but they refused to release them past beta
00:25<TheFirst>if you want their support then you do
00:25<warewolf>and they will support it
00:25<TheFirst>hell they provided fixes and then said in stall this but we cant support you
00:25<warewolf>you just don't want to pay (that much) for it. Same deal with microsoft.
00:26<TheFirst>not to mention when a ticket to them was marked as urgent it took them 4 days to respond
00:26<warewolf>if you install anything that isn't microsoft on a microsoft box, you're unsupported
00:26<warewolf>but oh yes they will accept large sums of money to get it working again.
00:27<TheFirst>never said microsoft was any better
00:27|-|Spads [] has quit [Ping timeout: 480 seconds]
00:28<TheFirst>what pissed me off about the situation is being forced to work around the confines of such crap just because they needed a support contract that in the end was entirely useless
00:28<warewolf>just saying that what you're upset about (and rightfully so) is industry standard practice
00:28<TheFirst>i completely realize that...
00:28[~]warewolf nods
00:29<TheFirst>and i'm not so much upset at rh for that either...i'd not expect more from them...what sucks is that the corp world is so clueless that they force that crap down your throats
00:29<TheFirst>and in doing so just make said crap more popular
00:29<warewolf>those support contracts spawned from old telco contracts
00:30<warewolf>"We'll install this nortel meridian phone system that runs OS/2 on a 486 for you, but you can't touch it. Only our engineers can touch it. If you break it, you get to keep both pieces."
00:30<TheFirst>i cant even say that it's like a few years ago when redhat was the leading distro for the home linux user (is it still....honestly dont know) and then to see it in a work environment was understandable at least ...keep what whoever set it up was familiar with
00:32<warewolf>it may very well be
00:32<warewolf>but then again, those sorts of numbers are extremely difficult to track
00:33<TheFirst>at my last job i founded funny that my boss worked at rh years back and hated them more than i did
00:34<warewolf>oh yes. The people who work for redhat are a particular breed. I don't know where they come from. I tend to try to stay away from them.
00:34<TheFirst>heh this guy was diehard debian but honestly he was a bit of a numskull in overall linux knowledge
00:35<warewolf>I don't care for debian because it trys too hard to be different
00:35<TheFirst>i dont like it on a desktop
00:35<warewolf>my gentoo experience was short-lived, but fun.
00:35<TheFirst>but for a server ... yah it's a bit outdated software but it's stable, updates dont break it and it works...
00:35<warewolf>I've heard that people adore debian on servers
00:35<TheFirst>gentoo i have no problems with on a desktop machine
00:36<TheFirst>i'd never run it on a server though
00:36<warewolf>I went the hardcore difficult route on gentoo: the hardened herd
00:36<warewolf>when your C compiler acts wonky, and it makes drastic changes to the way normal things work, interesting side-effects occur.
00:36<warewolf>like X not working.
00:37<TheFirst>gotta is it that you think deb tries to be diff?
00:37<encode>a decent package manager
00:37<warewolf>compare debian initscripts and module configuration stuff to everyone else
00:37<warewolf>goddamnit quit bashing RPM
00:37<warewolf>it's not the package manager that was the problem, it was bad packages
00:38<warewolf>and all package managers suffer from that problem
00:38<TheFirst>that may be but it's still a problem across rpm distros so it's easier to just generalize ;)
00:39<TheFirst>init scripts thing i cant judge...haven't used nondeb based distros in too long
00:39<TheFirst>except gentoo and that's just different in every way
00:39<warewolf>ugh christ
00:39<warewolf>I absolutely hated gentoo's initscripts
00:39<warewolf>if I could make it SysV-like I would maybe have kept gentoo
00:40<TheFirst>well now i tend to throw as much as i can into daemontools for it's restart if died ability so i avoid a lot of init scripting ...
00:41<warewolf>gah you've been bitten by the DJB bug
00:41[~]warewolf splashes you with holy water
00:41<warewolf>OUT EVIL SPIRIT
00:41<TheFirst>not by the philosophy...but the software just works ;)
00:41<TheFirst>and it is small which is important on my tiny xenode
00:41<warewolf>Yes! By assuming that an application has died, by it closing it's stdout.
00:41<warewolf>how intelligent
00:42<warewolf>I don't like in any way, shape or form, how daemontools is configured and used
00:43<warewolf>like how it MUST have some magic directory in /
00:43<warewolf>(or has that changed yet?)
00:43<Eman>i use daemontools 3.46... 4.x sucks majorly
00:44[~]warewolf firmly belives / is not to be shat on -- nothing more than bin, boot, sbin, usr, lib, opt, root, proc, var, export and home
00:44<encode>warewolf: i wasnt bashing rpm - its just that rpm is a package format, doesnt handle resolving dependencies like apt/aptitute does
00:44<warewolf>encode: I beat you to the punch is all :P
00:45<warewolf>you as well as I know that you were indirectly bashing it :P
00:45<encode>rpm and apt can't really be compared - they're two different things
00:45<warewolf>the last time I saw a feature line up, they were something like 90% similar
00:45<Eman>nothing in / ? you'd hate looking at my root.... 400+ files in / :p
00:45<encode>i guess i was - apt has treated me betterwith .dpkg than .rpm
00:46<encode>also, /me agrees wrt /
00:46<encode>files belong in folders
00:46<warewolf>and dev
00:46<encode>not in /
00:46<warewolf>I forgot dev in that list
00:46<warewolf>(dev and/or devices for solaris)
00:47<warewolf>I have a coworker who enjoys putting stuff in /ADMIN_NOTES and /SW_INSTALL
00:47<warewolf>and c:\ADMIN_NOTES and c:\SW_INSTALL
00:47<warewolf>(Yes, lets keep all the software and sysadmin notes on how on how to get/keep the box running on the box itself, so that when the box dies and is unrecoverable, we lose everything!)
00:48<Eman>so much crap hides in c:\ its quite a mess even if you dont put shit there on your own
00:48<encode>you mean, you dont back up those folders along with everything else?
00:48<encode>Eman: so true - its distressing
00:49<Eman>its like, why does windows insist on recreating 0byte files called io.sys and msdos.sys when you dont need them in nt?
00:49<encode>and autoexec.bat
00:49<encode>which is never executed
00:49<Eman>config.sys too
00:50<warewolf>please. I back up c:\Documents and Settings\warewolf\Temporary Internet Folders\7zyu0x6tnv.1st\{abafeedba53ba11-0000-cacafeedd0b33713511311345}
00:50<encode>:O you use internet explorer?
00:50[~]encode revokes warewolf's geek license
00:50[~]TheFirst watchs warewolf's credibility go down the toilet
00:50<warewolf>everything I download is in there!
00:51<Eman>ie has one use, downloading firefox or another browser
00:51<warewolf>shit from 1992 is in there!
00:51<warewolf>why are all these files zero bytes?!
00:51<encode>Eman: you missed the other use - exploiting systems
00:51<warewolf>holy shit there is a lot of porn in here!
00:51[~]warewolf ghasps
00:52<warewolf>goatse and tubgirl live there too!
00:52<Eman>why hide porn? i keed it in a directory called porn on my desktop
00:53<Eman>well, its not really on my desktop... its a shortcut to a share on my server, but w/e
00:53<warewolf>ok lets delete some stuff shall we .. select all .. delete SHARING VIOLATION: Unable to delete c:\Documents and Settings\warewolf\Tempoary Internet Folders\7z4jfhfroa.1kf\data\joe-random-folder. The file is in use.
00:54<warewolf>What? In use? Why is it in use? What is this a.exe in my task list?
00:54<warewolf>What do you mean my computer has to shut down because LSASS has crashed
00:55<encode>thats the annoying thing about windows, it locks files for no apparent reason
00:55<warewolf>who is this failed general, and why is he reading my drive c:?
00:55[~]Eman slaps warewolf for using the horrid crap known as windows xp... if you must use windows, use 2k, its not as good as linux, but better then xp
00:55<warewolf>Cool! My computer has its own doctor, named Watson. Maybe he's as smart as Sherlock.
00:57<Eman>oh, on the topic of stuff in c: that shouldnt belong there...
00:57<warewolf>autoexec.bat doesn't belong there anymore
00:57<Eman>php, mysql and openssl all like to be in c:\
00:57<warewolf>talk about depreciated into oblivion
00:57[~]warewolf shrugs
00:57<warewolf>never used any of those on a windows box
00:59<Eman>apache manages to bury itself 4 levels deep in program files
01:00<Eman>i have it all installed for testing stuff before it goes on my linode, i may be crazy, but i wont touch iis
03:14|-|afv-13 [] has joined #linode
03:24|-|streamcipher [] has quit [Quit: Leaving]
03:34|-|Karmatik [~gphreak@] has joined #linode
03:35|-|cmantito [] has quit [Quit: leaving]
03:35|-|Karmatik changed nick to cmantito
03:50|-|Dreameer3 changed nick to Dreamer3
03:54|-|Internat [] has quit [Quit: This computer has gone to sleep]
04:04<warewolf>caker- ping
04:04<warewolf>I know you're awake :)
04:06<warewolf>caker- unping, nevermind.
04:06<warewolf>anderiv- ping
04:16<cmantito>fstab is no one's friend.
04:19<afv-13>fstab is <3
04:19<cmantito>oh fuck
04:20<cmantito>crap crap crap
04:28|-|Dreameer3 [] has joined #linode
04:33|-|MarcelH [] has joined #linode
04:34|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
04:35|-|MarcelH [] has quit [Quit: ]
04:36|-|Marcel [] has joined #linode
05:03<warewolf>why oh why do I have afroman stuck in my head
05:15|-|afv-13 [] has quit [Read error: Connection reset by peer]
05:19|-|` [] has joined #linode
05:25|-|` [] has quit [Quit: Leaving]
05:33|-|afv-13 [] has joined #linode
05:34|-|FireSlash [] has joined #linode
06:04|-|FireSlash [] has quit [Quit: Leaving]
07:14<cmantito>my linode will NOT shut down.
07:15[~]encode pulls the power from cmantito's host
07:15<encode>problem solved :)
07:15<cmantito>yup, that worked :P
07:15<cmantito>(actually, something made it work, but don't ask me what)
07:16<cmantito>I ***really*** fucked something up =/
08:52|-|Spads [] has joined #linode
09:24|-|afv-13_ [] has joined #linode
09:24|-|afv-13_ [] has quit [Quit: ]
09:24|-|afv-13_ [] has joined #linode
09:28|-|Spads [] has quit [Ping timeout: 480 seconds]
09:31|-|afv-13 [] has quit [Ping timeout: 480 seconds]
10:09<cmantito>argh. alright, so I've just freshly installed the OS on my linode, wiped everything clean, set it all back up. got postfix, mysql, apache, sshd, dovecot all installed and configured and working, and every so often, I suddenly won't be able to login by ssh or through lish.
10:09<cmantito>ssh hangs after I enter my password, lish sits there after I enter my password and resets back to the login screen.
10:09<cmantito>(well, not lish, I can login to lish, but not through the console lish provides)
10:10<cmantito>if I have a session logged in already, I can keep using that, but there's nothing special in the logs.
10:11<cmantito>this is the 3rd time it's happened, and rebooting usually fixes it, but even rebooting is difficult because it doesn't want to shutdown properly. I have to send the shutdown once from the linode website and once from lish before it shuts down
10:11<cmantito>when it comes back up it's fine for a bit, repeat.
10:11|-|zmedico [] has quit [Remote host closed the connection]
10:12<@mikegrb>sending the shutdown twice isn't what's doing it
10:12<@mikegrb>there is a timout
10:12<cmantito>I figured that might be what it was.
10:12<cmantito>but either way, that's still not right.
10:12<@mikegrb>if your linode doesn't shutdown when a ctl+alt+del is sent, it sends an emergency sync then halt via sysreq
10:13<@mikegrb>2 minutes I believe is the timeout
10:13<@mikegrb>have you checked io_status via lish at the times this happens?
10:13<@mikegrb>I doubt it would be the problem
10:13<cmantito>my io is fine
10:13<@mikegrb>it's an odd problem
10:13<@mikegrb>naybe chris has seen it
10:13<cmantito>io_count=108636 io_rate=0 io_tokens=400000 token_refill=512 token_max=400000
10:14<cmantito>I have a session logged in.
10:14<cmantito>but I can't log anymore in.
10:14<@mikegrb>right now then?
10:14<@mikegrb>host52, right?
10:14<cmantito>right now.
10:14<cmantito>yup, host52
10:16<@mikegrb>apache and mysql and such continue to function normally?
10:17<@mikegrb>what does ps aux | grep syslog say?
10:17<cmantito>[root@redbaron log]# ps aux | grep syslog
10:17<cmantito>root 965 0.0 0.4 1552 532 pts/0 S+ 08:17 0:00 grep syslog
10:17<@mikegrb>Login timed out after 60 seconds.^M
10:18<@mikegrb>it sounds like it could be timing out waiting for syslog
10:18[~]cmantito goes to poke syslog
10:18<cmantito>which for some reason
10:18<cmantito>isn't running o.o
10:18<@mikegrb>we have this problem on the hosts when there are a few thrashers active at once
10:18<@mikegrb>that's why I asked about io first
10:18<@mikegrb>and your host looks great
10:19<@mikegrb>so seems like for some reason syslog is going south
10:19<cmantito>well, the RC script for it failed
10:19<cmantito>but running /usr/sbin/syslog-ng launched it
10:19<cmantito>and now it works
10:19<cmantito>so I'll futz with it and see what's up, but thanks :)
10:19<@mikegrb>no prob
10:19<@mikegrb>can you login now with it started?
10:20<cmantito>s'all working ok
10:34|-|jekil [~alessandr@] has joined #linode
10:38|-|FireSlash [] has joined #linode
10:42|-|jekil [~alessandr@] has quit [Ping timeout: 480 seconds]
10:52|-|FireSlash_ [] has joined #linode
10:56|-|FireSlash [] has quit [Ping timeout: 480 seconds]
11:23|-|G_w [tn@] has joined #linode
11:24|-|G_w changed nick to infrastructure
11:24|-|infrastructure [tn@] has left #linode []
11:38|-|FireSlash_ [] has quit [Quit: Leaving]
11:53|-|smohekey [~smohekey@] has quit [Ping timeout: 480 seconds]
12:07|-|FireSlash [] has joined #linode
12:10|-|Spads [] has joined #linode
13:10|-|FireSlash_ [] has joined #linode
13:10|-|FireSlash [] has quit [Ping timeout: 480 seconds]
14:06|-|FireSlash_ [] has quit [Read error: Connection reset by peer]
14:17|-|Spads [] has quit [Remote host closed the connection]
14:39<cmantito>anyone know a good way to tune down apache2's memory usage? it's eating like a whore at a free buffet! >.<
14:39<cmantito>I had apache 1.33 tuned down, but I had to upgrade to 2, and it seems to ignore the tunage for 1.
14:46|-|adamg [] has joined #linode
14:48|-|jekil [~alessandr@] has joined #linode
15:13<@caker>cmantito: httpd -V | grep MPM, then edit that section in your conf file
15:13<cmantito>ooh, thanks :)
15:17|-|Griswald [] has joined #linode
15:38|-|Spads [] has joined #linode
15:45|-|Battousai [] has quit [Quit: Leaving]
16:07[~]Eman drools
16:08|-|Battousai [] has joined #linode
16:26<@mikegrb>Eman: yes
16:30|-|Dream_3 [] has joined #linode
16:32<kriby>i've driven by that ballpark more times that i could count.. perhaps i should schedule a stop
16:35|-|Dreameer3 [] has quit [Ping timeout: 480 seconds]
17:02|-|Kiff [] has joined #linode
18:04|-|jekil [~alessandr@] has quit [Quit: Leaving]
18:06<weasel>hey mikegrb
18:07<weasel>or caker. ever seen clients trying to connect to port 0 TCP?
18:09<weasel>strange gets about 400kbit/s of that
18:22|-|Marcel [] has quit [Ping timeout: 480 seconds]
18:40<@caker>weasel: no, I've never seen that
18:40<weasel>looks like a puny attempt to DoS us
18:40<weasel>I sent mail to the whois contact, let's see what they say
18:41<@caker>from a quick google, looks like it might be used by certain port scanning techniques
18:42<weasel>port scanning port 0 for 3 hours?
18:42<warewolf>nmap's stealth scanning uses the same source port and can be told to use port zero
18:42<weasel>with tcp segments several kilobytes in size?
18:42<weasel>it's target port zero
18:42<warewolf>weasel- yeah, the large frames are unusual.
18:42<weasel><nitpick>it's a segment. frames are 2 layers down</>
18:43<warewolf>you know what I mean
18:59|-|Kiff [] has quit [Quit: Kiff]
19:55|-|silverblade [] has joined #linode
19:56|-|ximbiot [] has quit [Quit: Leaving]
19:56<silverblade>I've been told that MySQL wont perform very well on systems with < 128 MB RAM. Is this true? (im looking at the cheapest Linode plan). Also, how does linode compare with ?
19:59<@caker>silverblade: plenty of users on L80s run mysql just fine -- you might want/need to use the small memory config that comes with most mysql packages
20:01<@caker>silverblade: they use openvz (virtozzo) and we are using UML... with UML, you get your own kernel on up, whereas with openvz it's a shared kernel (and I think some of the more-fancy things don't work)
20:02<silverblade>Ah. See, I want to run a ShoutCast server but i have about half a years worth of hosting left with my current web host, so at the moment was thinking of using them for shoutcast only and then look to move the whole lot over to a single place later.
20:03<silverblade>Know if that's do-able with openvz/
20:06<silverblade>and also is there an alternative to mysql which might perform better with less ram?
20:06<warewolf>or a properly tuned mysql works fine
20:09<silverblade>ok, well thanks... i may not join for a little while but im considering linode for when my hosting runs out at the end of the year
20:10<warewolf>good luck, lots of people here would be hard pressed to find a better VPS provider.
20:17|-|javafueled [] has joined #linode
20:17|-|javafueled [] has left #linode []
20:19|-|silverblade9 [] has joined #linode
20:19<silverblade9>if anyone just said anything to me please repeat cos i got disconnected
20:21|-|silverblade [] has quit [Ping timeout: 480 seconds]
20:22<weasel>oh, it was actually your original client that quit now. sorry :)
20:22<weasel>Mon 03:09:05 <silverblade> ok, well thanks... i may not join for a little while but im considering linode for when my hosting runs out at the end of the year
20:22<weasel>Mon 03:10:13 <warewolf> good luck, lots of people here would be hard pressed to find a better VPS provider.
20:24|-|anderiv [~anderiv@] has quit [Quit: leaving]
20:29|-|anderiv [] has joined #linode
20:35|-|silverblade [] has joined #linode
20:35<silverblade>stupid unreliable connection :/
20:35<silverblade>i did say something after that... basically im a little hard up for cash at the moment so just going for cheapest option until my hosting runs out etc
20:36|-|silverblade9 [] has quit [Ping timeout: 480 seconds]
20:45<silverblade>anyway thanks for the info. will bear it in mind.
20:46|-|silverblade [] has left #linode []
21:09<encode>why do packages get kept back in apt?
21:25<tierra>apt-get dist-upgrade (new feature versions with new dependancies basically rather than bugfix/security releases or something else)
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
23:01|-|D[a]rkbeholder [] has joined #linode
23:01|-|darkbeholder [] has quit [Read error: Connection reset by peer]
23:36|-|afv-13_ [] has quit [Quit: leaving]
---Logclosed Mon Jun 12 00:00:12 2006