Back to Home / #linode / 2006 / 03 / Prev Day | Next Day
#linode IRC Logs for 2006-03-25

---Logopened Sat Mar 25 00:00:52 2006
00:10|-|darkbeholder [] has joined #linode
00:42|-|darkbeholder [] has quit [Read error: Connection reset by peer]
00:59|-|darkbeholder [] has joined #linode
01:10|-|Viza [] has quit [Quit: Leaving]
01:10|-|Viza [] has joined #linode
01:20<warewolf>holy crap that was quick
01:21<warewolf>I told google to nuke three false clicks on adsense and BAM they're gone from my account
01:22|-|spr [] has quit [Quit: Spoon!]
02:53|-|sec39 [] has quit [Read error: Connection reset by peer]
02:57|-|sec39 [] has joined #linode
03:30|-|myfnp [] has joined #linode
03:57|-|jekil [~alessandr@] has joined #linode
04:21|-|Marcel [] has joined #linode
04:23|-|MarcelH [] has quit [Ping timeout: 480 seconds]
04:34<alnr>need to do some rate-limiting, is connection tracking enabled in the kernel (debian)?
04:42|-|Marcel [] has quit [Ping timeout: 480 seconds]
04:44<vodka>cat /proc/net/ip_conntrack
04:46<alnr>tons of msgs in there so i guess so
04:47<alnr>also i think need to check for 'recent' module in iptables
04:47<vodka>best way to do that is just to try using it
04:48<vodka>if it fails, it's not present
04:49<alnr>was looking at , looks like it was a bit flaky for some people
04:49<alnr>even more than limiting connections, I might like to just 'slow' some of them down, idk if such a thing exists
04:50<vodka>if you have a 2.6 kernel, you could look in /proc/config.gz, where you'd see that iptables had the resent thingy built in
04:51<vodka>that file contains all the options the kernel was built with
04:52<alnr>no /proc/config.gz but i only have 2.24.xx so i guess thats why
05:00<alnr>hmm, traffic shaping
05:01<alnr>not for the faint of heart
06:45|-|FireSlash [] has joined #linode
07:00<warewolf>good lord
07:00<warewolf>at the 5 minute mark my linode is swamped for about 20 seconds while all the other linodes "check in" to RML :)
07:02|-|FireSlash [] has quit [Read error: Connection reset by peer]
07:09<vodka>curse pwgen and its unmemorizable passwords
07:12<vodka>ack! lighttpd allows you to use symlinks in the simple-host and evhost modules...
07:14|-|Marcel [] has joined #linode
07:14<vodka>shame... these would make life so much easier...
08:30|-|myfnp [] has quit [Quit: CGI:IRC (Session timeout)]
09:06|-|Marcel [] has quit [Ping timeout: 480 seconds]
09:57|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
09:59|-|Dreamer3 [] has joined #linode
10:14<chris>pwgen generates passwords that are purposefully easy to type
10:14<chris>I love it
10:57<Spads>excellent for shoulder-surfing
11:09|-|Marcel [] has joined #linode
11:11|-|spr [] has joined #linode
11:31|-|Marcel [] has quit [Quit: Leaving]
12:08<gpd>vodka: tried apg - pronounceable passwords?
12:23|-|vodka [] has quit [Quit: vodka]
12:25|-|vodka [] has joined #linode
12:26|-|vodka [] has quit [Quit: ]
12:49<warewolf>bad coder.
12:49[~]warewolf smacks somebody by the name 'chromatic'
12:55|-|Marcel [] has joined #linode
13:00<warewolf>sweet, I got a wiki working under mod_perl and HTML::Mason.
13:12<warewolf>wow, SQLite is really _really_ fast.
13:26|-|Viza [] has quit [Quit: Leaving]
13:27|-|Marcel [] has quit [Quit: Leaving]
13:28|-|Viza [] has joined #linode
13:31|-|alnr [] has quit [Quit: leaving]
13:44|-|vodka [] has joined #linode
14:01|-|Marcel [] has joined #linode
14:18|-|ghost [] has joined #linode
14:18<ghost>hey guys
14:18<ghost>whats u
14:19<ghost>anyone here?
14:21|-|ghost [] has quit [Quit: ]
14:21|-|darkbeholder [] has quit [Read error: Connection reset by peer]
14:22|-|darkbeholder [] has joined #linode
15:00|-|Marcel [] has quit [Quit: Leaving]
15:08|-|linbot [] has joined #linode
15:34|-|vodka [] has quit [Quit: vodka]
15:36|-|vodka [] has joined #linode
15:39|-|vodka [] has left #linode []
16:25|-|Spads [] has quit [Read error: Connection reset by peer]
16:26|-|Spads [] has joined #linode
17:06|-|FireSlash [] has joined #linode
17:38<taupehat>"What if we ignore your procedures? What if we say no? What will you do then? Continue to verbally attack us? To what end? To show that you are thankless dogs?"
17:38<taupehat>go theo!
17:55<taupehat>hey, caker or mikegrb... if you're around, there would seem to be host load issues on #15
17:55<taupehat>it's thrashing pretty hard, and I'm not doing it. traceroute looks good all the way in
18:02<@mikegrb>looks good to me
18:03<taupehat>naturally =]
18:03<taupehat>there's a corrolary to murphy's law that handles this situation rather well
18:04<taupehat>I encounter it all the time at work... I show up to fix something, only to find a flustered user saying something like "I swear it was doing that earlier! It seems fine now."
18:12|-|FireSlash [] has quit [Read error: Connection reset by peer]
19:01|-|jekil [~alessandr@] has quit [Quit: Leaving]
19:52|-|brandon20 [] has joined #linode
19:53|-|brandon20 [] has left #linode []
19:54|-|FireSlash [] has joined #linode
20:04<caker>what's the name of the debian package that finds the nearest/quickest mirror?
20:37<@mikegrb>caker: netselect is a generic tool
20:38<@mikegrb>apt-spy or netselect-apt are debian packages that will do just what you want
21:04|-|dc0e [] has quit [Quit: ↻ or ↓]
21:16|-|dc0e [] has joined #linode
21:25<fake>hosted gmail would be a lot cooler if they made the login/admin process easier
21:56<caker>huh .. in vi, n finds down, but N finds up .. learn something new everyday
21:57<caker>my vi abilities are impressive; only considering that I've never RTFM :)
21:57<bendy24>caker: hit d a bunch of times
21:57<caker>basically: ESC x4 after every operation
21:57<caker>bendy24: been there, done that :)
21:58<fake>what kind of linux master are you caker
21:58|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
21:59<caker>fake: the kind that can delete your filesystems
22:00|-|Dreamer3 [] has joined #linode
22:00<gpd>caker: do you know * and #? search f or b for word under cursor - quite useful
22:01<caker>gpd: nice :)
22:07|-|FireSlash [] has quit [Read error: Connection reset by peer]
22:08|-|Vaelor [] has quit [Quit: ]
22:12<encode>i knew about #, discovered it by accident
22:12<encode>by not about *
22:12<encode>thx gpd ;)
22:13<encode>didnt know about N either
22:13<caker>on the topic of deletes ... you can do d, then a number, then an arrow key, and it'll delete that many lines/chars in that direction
22:14<caker>often to clear a file (probably an easier way): d99999<down arrow>
22:14<encode>isnt it easier to echo "" > filename
22:14<encode>to clear a file?
22:15<caker>encode: even easier is just "> filename"
22:15<encode>i guess it would be
22:24|-|dc0e [] has quit [Remote host closed the connection]
22:33<warewolf>I have got to figure out what is causing people's result data sent to RML to be fucked up
22:33<warewolf>this is puzzling me
22:34<warewolf>where people's token refill value is 512, I'll get "5ok" or something else completely absurd
22:36<warewolf>5fi, fil, tim
22:36<warewolf>how the hell is 512 getting changed into those values
22:38|-|dc0e [] has joined #linode
22:38|-|Newsome [] has joined #linode
22:40<caker>warewolf: just certain people or random?
22:40<caker>and not every time, etc?
22:40<warewolf>fairly random, and no not every time
22:40<warewolf>but I did get a whole slew of them today
22:41<caker>are you logging the data before you do anything with it? perhaps it's on your end
22:41<warewolf>I'm beginning to think it's my linode getting swamped with the effective DDoS of the linodes checking in
22:41[~]caker shrugs
22:41<warewolf>the data would appear to be getting corrupted inside apache
22:42<warewolf>since SQLite appears to be really *really* blazing fast (and I'm not doing anything that really requires mysql) I may have to migrate over to SQLite
22:42<warewolf>hmm, shit.
22:42<Newsome>wow, the OOM killer is really running like crazy for me tonight.
22:42<warewolf>I do need mysql for the php forums
22:43<warewolf>and jabber.
22:43<warewolf>but I think everything else can use SQLite
22:44<caker>Newsome: legitimate OOM kills, or kernel bugginess?
22:45<caker>Newsome: which kernel?
22:45<Newsome>caker: 2.6.15-linode16
22:45<Newsome>I can't see anything it's killing. It just keeps spewing
22:46<Newsome>Total swap = 263160kB Free swap: 0kB
22:46<Newsome>oom-killer: gfp_mask=0xd0, order=2
22:46<Newsome>printk: 7328 messages suppressed.
22:47<warewolf>yeah you're out of memory
22:49<Newsome>caker, is it possible to send a 'sysrq t' or something? I can't tell why it's full
22:49<caker># uml_mconsole /linodes/sorenson/.uml/sorenson/mconsole sysrq t 2>/dev/null
22:49<caker>no output
22:49<caker>see anything in dmesg?
22:49<warewolf>caker- can that be done via lish?
22:49<Newsome>I can't even get into the box
22:49<warewolf>caker- I'm assuming no
22:49<caker>warewolf: it could be, sure
22:49<warewolf>Newsome: he means lish
22:50<caker>there's also /proc/sysrq-trigger
22:50<warewolf>caker- on second thought, sysctl .. hmm. actually it might not be that bad to have it, I can't think of some threat to the host OS, but there very well may be.
22:50<Newsome>no, nothing more in lish than just the every-5-second OOM
22:52<Newsome>how about sending 'sysctl f' once or twice, to see if it will _really_ kill something that's full...
22:52<warewolf>-bash-2.05b# echo t > sysrq-trigger
22:52<warewolf>-bash-2.05b# dmesg
22:52[~]warewolf is puzzled
22:52<warewolf>what is sysrq-trigger supposed to do?
22:53<caker>Newsome: sysrq f? (sysctl??)
22:53<Newsome>er, right
22:54<Newsome>dang. Not changing anything. I suppose I'll have to just reboot :(
22:54<caker>hmm, t on nova didn't do anything as well
22:54<caker>is sysrq broken in UML?
22:54<warewolf>it very well may be
22:54<Newsome>I thought it should work
22:54<caker>everybody: echo t > /proc/sysrq-trigger; dmesg <-- anything there?
22:55<caker>nova's on an oldish 2.4-um kernel, as well
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
23:04<warewolf>An error was detected while a client was attempting to update:
23:04<warewolf>illegal attempt to update using time 1143349200 when last update time is 1143349260 (minimum one second step)
23:04<warewolf>!math 1143349260-1143349200
23:05<warewolf>!calc 1143349260-1143349200
23:05<linbot>warewolf: 1,143,349,260 - 1,143,349,200 = 60
23:05[~]warewolf blinks
23:05<warewolf>how was it one minute in the past from the last update
23:05<caker>which host?
23:05<warewolf>host 24
23:05<warewolf>but again, random
23:06<caker>ok, I didn't touch that one
23:06<warewolf>and it isn't every update
23:06<warewolf>the last time that happend yes you did fsck the host RTC :)
23:07<warewolf>I get an email every 5 minutes per-client on that host when that happens :)
23:07<warewolf>so you will know REAL QUICK from me when that happens :)
23:08<warewolf>also, my math is wrong on the numbers on the rankings page
23:08<warewolf>my "average" across the entire time period (daily, weekly, monthly, yearly) in perl does not match the RRD's average
23:09<warewolf>how? I don't know.
23:09<warewolf>I retrieve all the values out of the RRD and average them out in perl
23:09<caker>RRD's is an average of the averages
23:09<caker>no clue
23:09<warewolf>averag io rate: 81
23:10<warewolf>that isn't what my 5m daily average in perl is on the ranking page
23:10<warewolf>it's 100 something or other
23:10<Newsome>caker: 'echo t > /proc/sysrq-trigger' worked fine for me (after reboot, at least)
23:10<warewolf>er, 24
23:18<warewolf>ugh god
23:18<warewolf>I need a new bed
23:19<warewolf>my back is killing me
23:20<warewolf>jesus are beds expensive
23:21<Newsome>Mohammed you're right, they are.
23:21[~]warewolf &
23:22[~]caker checks in a big-ass change to Linode .. hopes he didn't break anything
23:25|-|spr [] has quit [Quit: Spoon!]
23:25<warewolf>what'd you break?
23:27[~]mikegrb 's new chair sucks
23:27<@mikegrb>that's what I get for buying a chair that didn't havea display model
23:27<@mikegrb>maybe heidi will want it
23:30<warewolf>I'm about to spend $1000 on a new bed online w/o ever seeing it in real life
23:30<warewolf>should I?
23:38<Eman>craft matic adjustable II?
23:41<caker>sleepnumber system 9000?
23:46<caker>wow .. Sleep System 9000 actually exists. I thought I made up the 9000 part :)
23:52<warewolf>blackbash-2.05b# blackhole
23:52[~]warewolf hates people hammering his smtp server
23:52<warewolf>root 18898 1 0 Mar24 ? 00:00:02 sendmail: rejecting connections on daemon LOCALTLSMTA: 5 children, max 5
---Logclosed Sun Mar 26 00:00:41 2006