Back to Home / #linode / 2006 / 02 / Prev Day | Next Day
#linode IRC Logs for 2006-02-24

---Logopened Fri Feb 24 00:00:10 2006
01:47<linbot>New news from forums: Sticky: Offsite Backups HOW-TO in Linux Tips, Tricks, Tutorials <>
02:37|-|sab [] has quit [Quit: CGI:IRC (Session timeout)]
03:22|-|Shaun2222 [] has joined #linode
03:26|-|Shaun [] has quit [Ping timeout: 480 seconds]
03:27|-|Shaun2222 changed nick to Shaun
04:17|-|jimcooncat [] has joined #linode
04:43|-|jimcooncat [] has left #linode []
04:57|-|jekil [] has joined #linode
05:02|-|harshy [] has joined #linode
05:30<linbot>New news from forums: Importing a mysqldump with BLOBs in Linux, Apache, Mysql and PHP (LAMP) Forum <>
05:46|-|harshy [] has quit [Ping timeout: 480 seconds]
06:16|-|jekil [] has quit [Quit: Leaving]
07:00|-|linville [] has joined #linode
07:03|-|Redgore [~Redgore@] has joined #linode
08:42<linbot>New news from forums: Setting up the "Web Hosting" package. in Linux, Apache, Mysql and PHP (LAMP) Forum <>
08:44|-|jekil [~alessandr@] has joined #linode
10:41|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
10:42|-|Dreamer3 [] has joined #linode
11:05<linbot>New news from forums: Is Linode for me? in Sales Questions and Answers <>
11:08<kvandivo>linbot: no.. i don't think you probably need it
11:27|-|Dreamer3 [] has quit [Quit: Leaving]
12:29|-|FireSlash [] has joined #linode
12:38|-|Joe__ [] has joined #linode
12:38|-|Joe__ [] has left #linode []
13:02|-|FireSlash [] has quit [Quit: Leaving]
13:11|-|iggy [~iggy@] has quit [Quit: Lost terminal]
13:22|-|iggy [~iggy@] has joined #linode
14:23|-|FireSlash [] has joined #linode
14:31<FireSlash>Is host46 going nuts?
14:32<FireSlash>Go Go apt-get update
14:50|-|harshy [] has joined #linode
14:57<linbot>New news from forums: ssh: connect to host port 22: Connection refused in Linux Networking <>
15:29|-|VS_ChanLog [] has quit [Remote host closed the connection]
15:29|-|Beirdo [] has quit [Remote host closed the connection]
15:29|-|linbot [] has quit [Remote host closed the connection]
15:29|-|VS_ChanLog [] has joined #linode
15:29|-|encode [] has quit [Remote host closed the connection]
15:29|-|Beirdo [] has joined #linode
15:29|-|linbot [] has joined #linode
15:29|-|encode [] has joined #linode
15:45|-|Battousai [] has quit [Remote host closed the connection]
15:45|-|Battousai [] has joined #linode
15:54|-|alain [] has joined #linode
15:55<alain>hello all. I have been using the linode for nearly two years and I like it a lot. I only have one problem that might be the reason for me to leave the service
15:56<Battousai>i'm attracted to mikegrb too
15:56<alain>once every while, the io_status drops too low (probably because of httpd) and never gets back up again... or it takes too long to do so
15:56<alain>is there any way to prevent that?
15:56<Battousai>your tokens?
15:56<kvandivo>unless you have an extremely busy website you shouldn't be down that low because of traffic
15:57<kvandivo>are you positive that you aren't biting into swap?
15:57<alain>Swap: 256 39 217
15:58<kvandivo>well, you know the facts as well as anyone.. only thing i would mention is that Xen is "close" to coming out, and when it gets here that issue will likely go away
15:58<alain>to solve the problem, I just stop httpd for say... 5 minutes and allow the io_tokens to rise again
15:58<kvandivo>as best i recall, tests seemed to indicate that the io_token thing wasn't needed anymore on Xen.. someone else might correct me, though
15:59<alain>when io_token is high, the oscillations happens but are well tolerated. The problem is when the oscillations hit zero...
15:59<kvandivo>how big a plan do you have?
16:00<alain>i have the small one
16:00<alain>is it time to upgrade?
16:01<@mikegrb>alain: just reduce your ram usage
16:01<@mikegrb>tune apache and mysql, whatever is running
16:01|-|harshy [] has quit [Quit: Ex-Chat]
16:01<@mikegrb>most default config files assume a machine with lots of ram
16:01<Battousai>maybe even switch httpds if needed?
16:01<alain>I already have caker's configs for mysql and apache
16:01<Battousai>or cache your php a little bit
16:02<@mikegrb>do you have "skip-innodb" in the mysqld section of my.cnf?
16:03<@caker>alain: cat /proc/io_status please?
16:04<alain>io_count=1243567674 io_rate=0 io_tokens=362858 token_refill=512 token_max=400000
16:04<alain>now its running well
16:04|-|Stan [] has joined #linode
16:05<alain>io_rate is generally below 200
16:06<alain> is enabled in my.cnf
16:07<alain>maybe my apache is not configured properly? I can get about 65K hits per day on my pages
16:08<alain>here we go: io_count=1243745514 io_rate=1595 io_tokens=305001 token_refill=512 token_max=400000
16:08|-|spr [] has joined #linode
16:09<@mikegrb>yeah, and 103 MB in swap
16:09<alain>i am hitting zero soon. I will have to stop httpd again
16:09<@mikegrb>your average io_rate since boot is 70 which is quite high
16:09<alain>Swap: 256 106 150
16:09<kvandivo>yep.. that's your problem
16:10<@mikegrb>type "vmstat 1 5"
16:10<efudd>hrm. how do i see avg iorate since boot? :)
16:10<@mikegrb>and paste the output
16:10<@mikegrb>efudd: yes ;)
16:10<alain>procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
16:10<alain> r b swpd free buff cache si so bi bo in cs us sy id wa
16:10<alain> 0 0 107652 2840 1508 11984 2 2 1 1 0 0 4 1 94 0
16:10<alain> 0 0 107652 2700 1516 12104 0 0 120 356 232 29 8 1 91 0
16:10<alain> 0 0 107652 2700 1516 12104 0 0 0 0 102 14 0 0 100 0
16:10<alain> 0 0 107652 2700 1516 12104 0 0 0 0 102 15 0 0 100 0
16:10<alain> 0 0 107652 2700 1516 12104 0 0 0 0 102 14 0 0 100 0
16:10<efudd>I'm being dumb. :)
16:10<@mikegrb>well you are idle there
16:10<efudd>not really.
16:11<efudd>any swap is bad swap!
16:11<efudd>i don't want to do the math on uptime in seconds.
16:11[~]efudd sigh
16:11<kvandivo>the mere fact that the swap in use changed from 39 to 106 in between the two pastes tells me that is his problem
16:11<@mikegrb>it will almost always have some values for si and so on the first line
16:11<efudd>cat /proc/uptime
16:11<efudd>2073237.67 2059251.92
16:11<@mikegrb>kvandivo: ja
16:11<efudd>how convenient.
16:11<@mikegrb>efudd: first value
16:11<@mikegrb>second one is like total cpu seconds or something weird
16:12<@mikegrb>non idle cpu seconds I think
16:12<efudd>close enough that it matters little, but nod.
16:12|-|harshy [] has joined #linode
16:12<@caker>alain: apache 1 or 2?
16:12<efudd>in contrast, my "average made up mnemonic of io_rate since uptime" is 14, by comparison.
16:13<@caker>alain: which MPM? (I think: httpd -V | grep MPM or maybe -v)
16:13<alain>-D APACHE_MPM_DIR="server/mpm/prefork"
16:13<FireSlash>Someone might know: Is it normal for apache2 to have a boatload of spare processes running?
16:14<FireSlash>I just use it for SVN and theres got to be 10 running on top of 7 or so apache processes.
16:14<@caker>alain: ok, now edit your httpd conf file, and scroll down to the prefork section -- what are the values for MaxClients, etc?
16:14<@caker>FireSlash: follow along
16:15<alain>MaxClients 15
16:15|-|marcel [] has left #linode []
16:15<FireSlash>caker, I have been kind of, hence why I brought it up
16:15<@caker>alain: ok, and the other values?
16:15<alain>StartServers 4
16:15<alain>MinSpareServers 5
16:15<alain>MaxSpareServers 10
16:15<alain>MaxClients 15
16:15<alain>MaxRequestsPerChild 500
16:16<@caker>alain: I'd take MaxRequestsPerChild down to like 100 or so, and reduce MaxClients to 10
16:16<@caker>alain: maybe lower for MaxClients (rest will be queued)
16:16<@mikegrb>and minspare to like 2 and max spare to like 5
16:17<kvandivo>i use minspare 1, maxspare 3, startserv 1, maxlcients 8
16:17<alain>done. restarted.
16:18<alain>caker: do you think that that is the minimum or I can lower
16:18<alain>lower more?
16:19<alain>damn: io_count=1244374692 io_rate=441 io_tokens=-81 token_refill=512 token_max=400000
16:19|-|Stan [] has left #linode []
16:19<@caker>geez .. I suspect something else is eating IO
16:19<alain>I am now below zero... will never raise until I stop httpd
16:19<@caker>or maybe that's your swap unloading ...
16:20<alain>Swap: 256 87 169
16:20<@caker>alain: anything else eating memory (top, then "M" to sort my memory usage)
16:20<@caker>alain: dynamic pages?
16:20<@caker>alain: indexes on your mysql tables?
16:20<@mikegrb>yeah, he had 27mb swap after restarting apache
16:20<@mikegrb>and is back up to 87
16:21<@caker>alain: table scans are bad :)
16:21<alain>caker: i have always used phpnuke... I know it sucks, but today I migrated everything to drupal
16:21<@caker>alain: did you turn off all the unused apache modules?
16:21|-|harshy [] has quit [Ping timeout: 480 seconds]
16:22<alain>caker: probably not
16:22<alain>checking now
16:22|-|FireSlash_ [FireSlash@] has joined #linode
16:27<alain>it just doesnt stand still anymore... mmmmm
16:27<alain>i can loose 50K tokens in a matter of seconds
16:27<@caker>alain: if you'd like I can take a look ... log into Lish and then disconnect, I can get in that way
16:28|-|FireSlash [] has quit [Read error: Operation timed out]
16:28<alain> log into lish and then close the terminal?
16:28<@caker>alain: that or control-a d then exit
16:28<alain>ok thanks for your interest caker. Will log in now
16:29<efudd>m/sg caker SAY #linode Don't worry, I'll keep your data safe.
16:29<@caker>(after you log into your console)
16:29[~]efudd runs away
16:29<efudd>o/~ Obfuscation.. the king of Obfuscation...
16:29<FireSlash_>Hmm. made the changes and it didn't seemt o alter the number of processes >.>
16:30<alain>I just found this on the lish: __alloc_pages: 0-order allocation failed (gfp=0x1d2/0)
16:30<@caker>yeah, at some point you ran out of mem+swap
16:31<@caker>let me know once you've logged into your console and disconnected
16:31<alain>did that already
16:31<alain>hold on
16:32<alain>sorry i dint yet
16:33<alain>ok now I logged to the lish first and to my root account after
16:33<alain>earlier I only logged on the lish
16:34|-|harshy [] has joined #linode
16:34<@caker>alain: after you log in as root, type: control-a then d to get back to the Lish prompt, then "exit"
16:34<@caker>looks like disconnecting logs you out of the console
16:36<alain>nothing happens with control-a is that normal?
16:36<@caker>then d
16:37<@caker>ok, I'm in :)
16:38<alain>wow... someone is in from tty0 ! eheheh
16:38<Spads>rooooot has LANDED
16:41|-|linville [] has quit [Quit: Leaving]
16:41<alain>i feel a little worried now... being examinated by the guru is not an easy moment
16:42<@caker>alain: mind making some noise on your websites?
16:42<alain>noise? what do you mean?
16:42<alain>stop apache?
16:42<@caker>make some hits in teh browser
16:42<alain>go ahead
16:42<alain>I am doing that already
16:43<@caker>ok, good
16:43<@caker># cat /proc/swaps
16:43<@caker>Filename Type Size Used Priority
16:43<@caker>/dev/ubdb partition 263160 30368 -1
16:44<@caker>/dev/ubdb partition 263160 52056 -1
16:44<@caker>looks a lot better
16:44<@caker>I cut the apache values in half
16:44<FireSlash_>Filename Type Size Used Priority
16:44<FireSlash_>/dev/ubdb partition 263160 13952 -1
16:44<FireSlash_>With apache and apache2
16:44<FireSlash_>Do I win a cookie? :)
16:44<alain>the swap is being less used?
16:45<@caker>yeah -- you still have a lot of apache modules that are good candidates for removal
16:45<alain>go ahead. You're the boss
16:45<@caker>modules/ <-- ??
16:45<@caker>nah, that's your job -- I have no clue what you're using
16:45<@caker>modules/ <-- ?? wtf are these (I've never really looked, tbh)
16:46<alain>i didnt add them
16:46<alain>they were there
16:46<alain>do I have to get worried?
16:46<@caker>well, your swap usage is hanging at about 40M so far .. I'd say we've get it under control
16:47[~]caker logs out
16:47<alain>ok caker thanks so much for your help. Linode RULES
16:51<FireSlash_>Also, thanks for the help. Shaved down apache and apache2 on my side, got some memory freed up :)
16:51<tierra>wow, that's a first I've seen caker go to that length to help
16:53<tierra>not to say he doesn't help, but that's going quite out of the way
16:55<FireSlash_>Well, he helped multiple people out in the process, soo.... :)
16:59<alain>caker: what have you done? I never seen the linode so FAST!!! wow
16:59<@caker>alain: heh .. just reduce those apache conf settings
17:00<kvandivo>check for the existence of the /etc/caker_rules file
17:01<FireSlash_>kvandivo, I was thinking more along the lines of ps -A | grep caker
17:01|-|sec39 [] has quit [Quit: ]
17:02|-|sec39 [] has joined #linode
17:20|-|harshy [] has quit [Ping timeout: 480 seconds]
17:21|-|FireSlash_ [FireSlash@] has quit [Ping timeout: 480 seconds]
17:38|-|FireSlash [] has joined #linode
17:42<taupehat>tierra: you weren't here to watch caker (and mikegrb) laugh and help me when I tried to install Open X-Change server on my /64
17:43<taupehat>something about not having IO tokens for a few years
17:52|-|jekil [~alessandr@] has quit [Quit: Leaving]
17:57|-|spr [] has quit [Quit: Spoon!]
18:05<alain>thanks to everyone for the help
18:05<alain>linode rules!
18:05|-|alain [] has quit [Quit: Sto andando via]
18:06<linbot>In 1972, a crack commando unit was sent to prison by a military court for a crime they didn't commit. They promptly escaped from a maximum security stockade to the Los Angeles underground. Today, still wanted by the government, they survive as soldiers of fortune. If you have a problem, if no-one else can help, and if you can find them, maybe you can hire the A-Team.
18:07[~]caker downloads Web 2.0
18:29[~]gpd <3 linbot
19:25<linbot>Attitude Determines Altitude
20:20|-|cout [] has quit [Ping timeout: 480 seconds]
21:16|-|anderiv [] has joined #linode
21:30|-|mark [] has joined #linode
21:30|-|FireSlash_ [] has joined #linode
21:32<mark>Question about passwords: I am 10 minutes new to Linode. I can log in to my Linode account but my password does not work when I try to log to my root account. What password should I use to log in as root?
21:37|-|FireSlash [] has quit [Ping timeout: 480 seconds]
21:40<@caker>mark: the password you entered when you used the Distro Wizard
22:00<mark>OK, thanks, i've rebuild the config using the Wizard and I'm in now...
22:01<mark>I guess my pw was a little to fancy...
22:01|-|FireSlash_ [] has quit [Quit: Leaving]
22:05|-|grouchhouse [] has joined #linode
22:22|-|grouchhouse [] has quit [Quit: grouchhouse]
22:30|-|FireSlash [] has joined #linode
22:41|-|FireSlash [] has quit [Read error: Connection reset by peer]
22:49|-|Newsome [] has joined #linode
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
23:23<mark>I have a basic debian distro running. I want to use vnc but it appears to not be installed. Then I run 'apt-get install vncserver' and it appears to install but when I try to launch with '# vncserver' there appears to be no application to launch. Also entering 'man vncserver' returns an error indicating there are no man pages. Any suggestions as to how to install vnc? Thanks.
---Logclosed Sat Feb 25 00:00:04 2006