1162166416 M * doener (or those of the vserver+grsec merge) 1162166416 M * daniel_hozac we do that too with the accounting ;) 1162166427 M * daniel_hozac (i.e. the lock issues and such) 1162166430 M * Radiance not that i'm an expert but spender mentioned it at several occassions when ppl kept repeating such warnings hehe 1162166621 M * pzYsTorM daniel_hozac: of course, suphp is a must-have. but it doesnt help you if someone installs via phpShell a rootkit and tries to get root-rights. or sees, that a process is running which can be compromised. 1162166651 M * pzYsTorM Radiance: the buggy app would be /lib/util-vserver/secure-mount ... 1162166670 M * daniel_hozac pzYsTorM: if the process can be compromised, there are other ways to detect it. 1162166731 M * daniel_hozac pzYsTorM: and how does process hiding help you to protect against rootkits? 1162166755 M * pzYsTorM running a chkrootkit every minute ;) 1162166891 M * pzYsTorM perhaps: "uncertainty which processes will detect me / my intrusion" or: "disinterest, because the security is to high for me" 1162166932 M * daniel_hozac surely anyone who tries to break in would consider high security a challenge and thus would get even more interested? 1162167053 M * pzYsTorM would he get even much more interested, if he sees all those processes and could pre-estimate what damage he could create? 1162167081 M * daniel_hozac well, process-hiding is pretty obvious. 1162167125 M * daniel_hozac and if you've got something to hide... 1162167279 M * daniel_hozac if you were to come across a wall in the middle of the woods, with a big "keep out" sign, wouldn't you get interested to see what's on the other side? 1162167407 M * pzYsTorM then the only question would be: how high is the wall? 1162167429 M * daniel_hozac why climb it if you can find a door? 1162167436 M * daniel_hozac alternatively, make a door 1162167581 M * trippeh All suphp setups I've seen to date gives the webapps a completely writable webroot, which is a massive security WTF. 1162167616 M * ntrs_ daniel_hozac, what is the best/official way to quickly stop and start all guests on a host? 1162167663 M * doener trippeh: yeah, you usually even end up as a "real" user, whose .bashrc you could manipulate to sneak in lots of stuff 1162167690 M * trippeh Thats is also a good point 1162167792 M * ntrs_ daniel_hozac, do the tools provide something for that? 1162167832 M * trippeh I wish there were a better solution to it, so far using suphp have only made more work restoring backups of defaced sites and hampered stability. The isolation is nice though, but I'm not sure its worth it for shared hosting setups. 1162167841 Q * matled Remote host closed the connection 1162167846 M * trippeh hampered scalability even 1162167853 M * doener ntrs_: ls /etc/vservers | xargs -n1 basename | xargs -n1 -i vserver {} stop 1162167869 M * doener that should probably work 1162167917 M * ntrs_ doener, I do have a script that starts and/or stops the vservers, but I was interested if there was something in the tools that would do that. 1162167945 M * doener not yet AFAIK, only the mark stuff with which you can start/stop a class of vservers 1162167962 M * ntrs_ ok thanks 1162167986 M * trippeh And grsec is written by a guy who openly said he would sell 0day exploits, so thats no go for me either. Massive conflict of interest. 1162168031 M * Aiken I was told this to stop all vservers /usr/lib*/util-vserver/start-vservers --stop --all 1162168038 M * trippeh I wished I bookmarked a reference for that though ;) 1162168045 M * trippeh wish, even 1162168057 M * doener Aiken: IIRC that is/was broken 1162168108 M * daniel_hozac doener: fixed in 0.30.211. 1162168116 M * doener ok 1162168117 M * daniel_hozac (and in most distro packages for 0.30.210) 1162168119 M * Aiken I have not had a bad shutdown since I started using it 1162168135 M * ntrs_ daniel_hozac, thanks 1162168182 M * daniel_hozac np. there's no restart though, just start and stop. 1162168202 J * Hacka ~Hacka@tor-irc.dnsbl.oftc.net 1162168237 M * trippeh Most of what people use grsec for seems to just be security by obscurity anyway. 1162168265 M * Hacka NIGGERS SMELL BAD 1162168290 F * ChanServ +o doener 1162168293 M * Hacka I HATE NIGGERS, THEY SMELL BAD 1162168297 M * daniel_hozac Hacka: what makes you say that? 1162168299 M * Hacka lol fuck u doener 1162168302 M * Hacka THEY SMELL SOOO BAD 1162168308 M * Hacka OOOOHH, THE STENCH! 1162168325 F * doener +b *!*Hacka@*.dnsbl.oftc.net 1162168325 K Hacka doener go wash yourself then 1162168333 M * daniel_hozac ehhe. 1162168340 F * doener -o doener 1162168410 M * daniel_hozac ntrs_: if you just want to restart running guests, something like vsomething vserver --all -- condrestart should work too. 1162168423 M * daniel_hozac (but will again require 0.30.211, as condrestart was broken) 1162168527 M * ntrs_ what is vsomething? 1162168539 M * daniel_hozac Execute * foreach vserver. 1162168565 M * ntrs_ wow, didn't know about that one. 1162168610 M * daniel_hozac rather handy little util. 1162168622 M * ntrs_ will condrestart do an actual restart of the guests? 1162168632 M * daniel_hozac it will do a vserver ... restart, if it's already running. 1162168650 M * daniel_hozac otherwise it won't do anything. 1162168682 M * ntrs_ ok, is that the fastest way to restart all the guests on the host? 1162168716 M * daniel_hozac if you don't have any non-running guests, start-vservers is probably faster as it can run several startups in parallell. 1162168723 M * daniel_hozac (and stops) 1162168729 M * daniel_hozac but it requires two steps instead. 1162168759 M * ntrs_ if all are running? 1162168769 M * daniel_hozac right. 1162168782 M * daniel_hozac otherwise it would start guests that weren't previously running 1162168794 M * daniel_hozac (as expected, since you're telling it to start all of them) 1162168814 M * ntrs_ so to stop or start all of them, running and non-running start-vservers is faster, right? 1162168828 M * daniel_hozac yep. 1162170025 J * brcc bruce@i.am.someasshole.com 1162170058 M * brcc without limiting VM(AS), is it possible for a vserver to crash the whole machine ?\ 1162170058 M * brcc \ 1162171222 J * s0undt3ch_ ~s0undt3ch@81.193.56.198 1162171645 Q * s0undt3ch Ping timeout: 480 seconds 1162171645 N * s0undt3ch_ s0undt3ch 1162171687 J * mire_ ~mire@233-167-222-85.adsl.verat.net 1162176925 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1162176935 J * ensc ~irc-ensc@p54B4D1CE.dip.t-dialin.net 1162177514 J * Aiken_ ~james@tooax6-212.dialup.optusnet.com.au 1162177515 Q * Aiken Read error: Connection reset by peer 1162180032 Q * ms_ Ping timeout: 480 seconds 1162180647 Q * ntrs_ Remote host closed the connection 1162182823 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1162185565 Q * _node Ping timeout: 480 seconds 1162186761 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162189431 Q * AndrewLee Quit: leaving 1162189900 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1162189924 J * besonen_ ~besonen@dsl-db.pacinfo.com 1162190295 Q * besonen Ping timeout: 480 seconds 1162190960 Q * besonen_ Ping timeout: 480 seconds 1162191379 Q * Piet Remote host closed the connection 1162191389 J * Piet hiddenserv@tor.noreply.org 1162191405 J * besonen ~besonen@dsl-db.pacinfo.com 1162192013 Q * besonen Ping timeout: 480 seconds 1162192059 J * matled ~matled@85.131.246.184 1162192148 J * besonen ~besonen@dsl-db.pacinfo.com 1162192598 Q * pzYsTorM Ping timeout: 480 seconds 1162192685 Q * besonen Ping timeout: 480 seconds 1162192751 J * besonen ~besonen@dsl-db.pacinfo.com 1162193028 Q * Piet Quit: Piet 1162194341 J * dna_ ~naucki@111-250-dsl.kielnet.net 1162195299 J * AndrewLee ~andrew@140.109.13.50 1162195458 Q * Aiken_ Quit: Leaving 1162196138 Q * phil_ Quit: leaving 1162196628 Q * dna_ Quit: Verlassend 1162197128 M * Borg- Radiance: on what Kernel version you had load issue? 1162197132 M * Borg- s/issue/issues/ 1162197492 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at 1162197598 Q * shedi Quit: Leaving 1162197834 M * renihs morning 1162201691 Q * AndrewLee Quit: leaving 1162201709 J * AndrewLee ~andrew@tnlug.linux.org.tw 1162201852 Q * AndrewLee 1162201856 J * AndrewLee ~andrew@tnlug.linux.org.tw 1162202192 Q * ms_ Ping timeout: 480 seconds 1162203051 Q * AndrewLee Quit: leaving 1162203065 J * AndrewLee ~andrew@tnlug.linux.org.tw 1162203097 Q * AndrewLee 1162203110 J * AndrewLee ~andrew@tnlug.linux.org.tw 1162205344 J * Sla ~sl@83.229.149.151 1162205359 M * Sla gi 1162205362 M * Sla hi 1162205404 M * Sla doener: seems, that apache hangs before mysql. 1162205862 M * Sla is anybody here? 1162206060 J * the_hydra ~a_mulyadi@202.59.168.29 1162206898 Q * the_hydra 1162207076 M * Sla ? 1162207794 J * MrX ~urk@219.95.238.69 1162208104 M * Sla hi! i've got the problem, when apache refuse to server under vserver. processes are runnng, establishing connections, but no answer from the server... 1162208434 M * goblin refuse to server? 1162208465 M * goblin how did you configure the network on the vserver and what Listen directive have you given to apache? 1162208743 Q * michal` Ping timeout: 480 seconds 1162208782 M * Sla hm... standart network configuration, as if ifconfig performs. 1162208798 M * Sla apache uses standart config, no sendfile. 1162208818 M * Sla it listens on separate addr, not 127 or so. 1162208917 M * renihs there is no lo in vserver 1162208923 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162208954 M * Loki|muh isn't sendfile in standard apache config? ;) 1162208973 M * Loki|muh Default:EnableSendfile On for apache2.0 1162208978 M * Sla situation is strange. httpd, mysql +php. so, when started, everything works OK, but later - I can see only processes of httpd and mysql with no action from them. also, there is a situation, when some of apache's virt. servers are working, some of them (most loaded) produce no answer. 1162208996 M * Sla sendfile switched off. 1162209013 M * Loki|muh ah 1162209035 M * derjohn Sla, is you storage on NFS? 1162209036 J * michal` ~michal@www.rsbac.org 1162209202 M * Sla I'm using hw partitions with fs on them. 1162210475 M * Sla ? 1162210851 M * derjohn Sla, does apache start with an "file not found" ? 1162210857 M * derjohn (on the rootdir= 1162210858 M * derjohn ) 1162211236 M * Sla no, seems, not 1162211248 M * Sla which file, if so? 1162211348 M * daniel_hozac Sla: have you placed any limits on your guest? 1162211763 M * Sla increased nofiles, nproc. (3900, 1000) 1162211770 M * Sla nothing else. 1162211792 M * Sla hanging just now. 1162211937 M * Sla and load average is very low. 1162211962 M * daniel_hozac are you hitting the nofile limit? 1162212301 M * Sla yes. 1162212518 M * daniel_hozac so you might want to raise that further then. 1162212556 M * Sla statistics says, that there are only 1700 file open. 1162212663 M * daniel_hozac so how could you be hitting it? 1162212721 M * Sla look at paste.linux-vserver.org for my report. 1162212769 M * daniel_hozac so you're not hitting it at all. 1162212804 M * Sla seems so. 1162212810 M * Sla any idea? 1162212895 M * Sla previous version (kernel 2.6.11.5 with vs 1.9.5) works fine. 2.6.18.1 and vs2.1.1 rc 42 hangs. 1162212993 M * derjohn Sla, take rc45 any any case 1162213001 M * derjohn are you on Debian ? 1162213098 M * Sla not. slack. 1162213113 M * Sla rc45... I see rc44 only. 1162213126 M * daniel_hozac then you're not looking in the right place :) 1162213150 M * Sla what is right? :) 1162213161 M * derjohn vserver.13thfloor.at/Experimental 1162213209 M * Sla oh! old one! :) 1162213876 M * Sla so, will check soon... several minutes... 1162215305 M * Sla i see, that the number of connections to /tmp/mysql.sock is not increasing all the time in rc45. for now, the server is working (yet). will keep it checking. 1162215710 M * derjohn Sla, how do you check the numer of connections ? 1162216290 M * Sla netstat 1162216421 M * Sla so, it hangs now. some servers are active, but most of them not given an answer. 1162216453 M * Sla the number of connections to mysql.sock increasing. 1162216457 M * derjohn there was a mysql problem reported by ntrs . but the issue was solved with early rc4x. 1162216461 M * derjohn (AFAIK) 1162216571 M * Sla hm... 1162216622 M * Sla I do not know, why old configuration is working, and the new one hangs. 1162216646 J * s0undt3ch_ ~s0undt3ch@bl4-63-208.dsl.telepac.pt 1162216715 M * derjohn Sla, I dont have problem with mysql so far. 1162216731 M * derjohn did you upgrade more than only the kernel ? 1162216766 M * Sla hm... only kernel/mysql/php/apache updated. 1162216864 M * derjohn Sla, maybe a mysql problem with threads ? 1162216872 M * Sla glibc 2.3.4 1162216895 M * Sla mysql and threads... why now, why not with 2.6.16 kernel? 1162217014 M * Sla it is running now in 115 threads 1162217031 M * Sla sorry, 114 1162217057 M * derjohn 115 ??? 1162217067 M * derjohn wow, that server is loaded .... 1162217072 Q * s0undt3ch Ping timeout: 480 seconds 1162217072 N * s0undt3ch_ s0undt3ch 1162217105 M * derjohn there might be a difference with the kind of threading slack uses. 1162217126 M * derjohn But I am really no expert with native vs. posix threads ... sry ! 1162217146 M * derjohn (or whatever they are called ... ) 1162217191 M * Sla pthreads, I think, cannot cause that. 1162217498 M * derjohn but 115 threads ? wow .... is the machine very loaded? did you limit nproc in that guest ? 1162217527 M * derjohn Sla, did you check /proc/virtual//likmis on the host ? 1162217724 M * Sla nproc limit..... is not good. 1162217764 M * derjohn Sla, but obivously not the cause of the "hanging only thing"! 1162217773 M * Sla /proc/virtual//likmis ? i see no such file.... 1162217821 M * derjohn *limits 1162217823 M * doener Sla: could you provide the output of "show variables;" and "show status;"? 1162217826 M * doener (mysql) 1162217994 M * Sla yes, I can. just rebooted.... 1162218431 M * Sla still working... waiting... 1162218529 N * Bertl_zZ Bertl 1162218533 M * Bertl morning folks! 1162218666 M * renihs morning Bertl 1162219004 A * derjohn away now. Sla: good luck and pls report ! 1162219108 J * s0undt3ch_ ~s0undt3ch@81.193.60.4 1162219527 Q * s0undt3ch Ping timeout: 480 seconds 1162219527 N * s0undt3ch_ s0undt3ch 1162219766 M * Sla doener: so, hanging. logs are in paste.linux-vserver.org 1162219788 M * Sla I will come back in an hour. 1162219792 Q * Sla Quit: Leaving 1162219820 M * harry hehe... "i pasted something on pastebin.com" check it out! ;) 1162219853 M * Bertl harry: well, it's quite easy if you add your _name_ :) 1162219869 M * harry Bertl: i know... it's just kinda funny to read :) 1162219890 M * harry i put my website online, check it out, i'm off now, will be back later-style 1162219891 M * harry ;) 1162219923 M * harry luckily the vserver pastebin is not a total wilderness as is the restof the internet :) 1162219954 M * Bertl yeah, btw, did you read about the stack issue with grsec/vserver? 1162219984 M * harry ow... no 1162219997 M * harry been away all weekend and so... 1162220030 M * harry where was the discussion? on irc? 1162220079 M * Bertl http://irc.13thfloor.at/LOG/2006-10/LOG_2006-10-29.txt (near the bottom) 1162220087 M * doener damn, 2fast4me 1162220103 M * doener had just looked at todays logs to see that it was the wrong one ;) 1162220120 M * Bertl :) you still have the mysql stuff 2reply2 :) 1162220347 A * doener doesn't like answering to the void 1162220378 M * Bertl yeah, what I meant is, you still got roughly an hour to prepare :) 1162220380 M * harry very odd... why does it want to allocate 128MB for a stack??? 1162220390 M * doener it probably doesn't 1162220394 M * Bertl read carefully, it keeps increasing ... 1162220394 M * harry normal stacks ARE limited to 8MB 1162220401 M * harry yeah... i'll read on 1162220402 M * doener but some counter or whatever is not reset 1162220405 A * harry will chheck some code 1162220453 M * doener Bertl: what I can tell so far is that there are some probably not-so-optimized queries at work ;) 1162220500 M * Bertl hehe is that supposed to 'hang' mysql? 1162220516 M * doener but I see no reason for a "lock up"... all threads are in use, but there are still 449 threads available for creation 1162220638 M * doener well, it _can_ hang the box if a query select a gazillion of rows from php... php's memory limit does not apply to that, it will happily store all rows and if you run mod_php, that will keep your apache process using huge amounts of ram 1162220669 M * doener (with buffered queries that is) 1162220692 M * doener but it's not mysql's fault in that case 1162220711 M * Bertl i.c. 1162220830 M * doener there are fd's left, connections left... my best bet atm is that mysql is so working with temp tables so hard, that it has no time to accept connections and the backlog ran full 1162220861 M * doener especially as the connection reserved for the super user connected fine 1162220956 M * doener hm, the "but that doesn't make sense to me either" part got lost somewhere 1162220957 M * harry it's just a logging feature of grsec it seems... you can easily disable that 1162220965 M * harry just logs... nothing more 1162220975 M * harry doesn't involve grsec code at first sight 1162221017 M * Bertl well, it states something about 'denied' 1162221090 M * harry indeed 1162221104 A * harry will look it up further 1162221216 M * harry you can easily disable logging of that stuff by sysctl -w kernel.grsecurity.resource_logging=0 1162221240 J * Piet hiddenserv@tor.noreply.org 1162221242 M * harry that will disable the logging... because that's all that the feature does... 1162221323 M * Bertl okay, but in this case, I think you should report it upstream 1162221355 M * Bertl it doesn't sound good to me when a 'logging' feature (which seems broken) 'claims' that it denied something 1162221390 A * harry looks where the logging function gets called 1162221730 M * harry i really need a trace on that one... 1162222008 M * harry bleh! => 0xc0ffee! 1162222449 M * harry Bertl: you want the diff on that error ? 1162222459 M * harry it's in mm/mmap.c 1162222468 M * harry original: /* Stack limit test */ 1162222468 M * harry if (size > rlim[RLIMIT_STACK].rlim_cur) 1162222468 M * harry return -ENOMEM; 1162222477 M * harry grsec one: 1162222478 M * harry /* Stack limit test */ 1162222478 M * harry gr_learn_resource(current, RLIMIT_STACK, size, 1); 1162222478 M * harry if (size > rlim[RLIMIT_STACK].rlim_cur) 1162222478 M * harry return -ENOMEM; 1162222509 M * harry it just logs a rule, before returning a -ENOMEM 1162222621 M * Bertl gr_learn_resource() doesn't sound like gr_log_overlimit() :) 1162222650 M * Bertl I'd assume it accounts whatever size is (unconditionally) 1162222675 M * Bertl but some coresponding gr_forget_resource()? is missing? 1162222700 M * Bertl note: I have no idea about grsec here, just guessing from the names :) 1162222838 M * harry hehe 1162222843 M * harry so i notice ;) 1162223199 M * transacid shit. i made a very very big mistake. i wanted to delete a vserver directory without shutting the vserver down. now i have a "ghost" vserver. is there any possibility to get it out of proc without rebooting? 1162223202 J * dna_ ~naucki@233-195-dsl.kielnet.net 1162223254 M * daniel_hozac vkill --xid -s 15; sleep 2; vkill --xid -s 9 1162223323 M * transacid ahh great, thx a lot daniel_hozac 1162223379 M * transacid but with that prob i figured out that there is a delete parameter ;) 1162223381 Q * [PUPPETS]Gonzo Remote host closed the connection 1162223438 J * shedi ~siggi@inferno.lhi.is 1162223661 J * [PUPPETS]Gonzo gonzo@langweiligneutral.deswahnsinns.de 1162223715 Q * [PUPPETS]Gonzo Remote host closed the connection 1162223750 M * doener Bertl: it really does just log the denial. It duplicates the check and logs it 2 level down in the call tree 1162223773 M * harry uh 1162223774 M * harry u 1162224033 J * [PUPPETS]Gonzo arrakaij@langweiligneutral.deswahnsinns.de 1162224125 M * Bertl doener: so be it ... leaves us with the question, why is the stack size increasing? 1162224217 Q * ms_ Ping timeout: 480 seconds 1162224438 J * s0undt3ch_ ~s0undt3ch@bl4-62-147.dsl.telepac.pt 1162224832 Q * s0undt3ch Read error: Operation timed out 1162224832 N * s0undt3ch_ s0undt3ch 1162225784 J * eyck_ eyck@ghost.anime.pl 1162226558 J * stefani ~stefani@tsipoor.banerian.org 1162227124 M * Bertl welcome stefani! hey eyck_! 1162227147 M * stefani salut 1162227839 J * sla 1006@leo.sinp.msu.ru 1162228241 J * s0undt3ch_ ~s0undt3ch@bl9-225-45.dsl.telepac.pt 1162228241 Q * s0undt3ch Read error: Connection reset by peer 1162228250 N * s0undt3ch_ s0undt3ch 1162228282 Q * hardwire Ping timeout: 480 seconds 1162228634 J * bonbons ~bonbons@83.222.36.111 1162228742 M * sla ? 1162228745 Q * SMuZZ Ping timeout: 480 seconds 1162228755 M * Bertl hey sla! 1162228769 M * sla hey! 1162228890 M * Bertl sla: you are the one with the mysql issues, right? 1162228948 M * sla ye 1162228976 M * Bertl doener said something about excessive queries and such 1162229032 M * doener sla: no real conclusion yet... what's the cpu load like when mysql "hangs"? 1162229041 M * sla strange behaviour. mysql or apache hangs after the period of activity. 1162229049 M * sla CPU load is about 10% 1162229057 M * doener 90% idle? 1162229061 M * sla yes. 1162229080 M * Bertl any limit hits shown? 1162229083 M * doener hm, kills the temp table activity idea... that would involve lots of %wa 1162229124 M * sla no one hit. 1162229134 M * sla and nproc was max at 457 1162229151 M * Bertl I know that is a stupid question, but, any logs from mysql? 1162229199 M * sla no, it thinks that everything is ok. status was posted at paste.li**** 1162229231 M * Bertl okay, how long does it take to 'hang'? 1162229256 M * sla hard to say, from 5 minutes up to 30 min. 1162229276 M * Bertl so relatively fast ... 1162229296 M * sla i'm looking at the mysql.sock connections via netstat. 1162229307 M * Bertl is it possible to attach strace to the processes somehow? 1162229309 M * sla whenthe number increases - hangs. 1162229333 M * Bertl so there is a socket limit which causes the hang? 1162229345 M * sla hm.... strace... it is running, if I put it to strace, it can go a good way. 1162229361 M * Bertl you can attach when that happens 1162229369 M * sla i set no limits. maybe, in rlimits? 1162229392 M * Bertl could be, but unlikely, I thought more of a connection limit in mysql 1162229400 M * sla how to attach? I'm not using strace often. 1162229420 M * sla connection limit is not overrunned. 1162229421 M * Bertl something like: strace -fF -o mysql.trace -p 1162229492 J * besonen_ ~besonen@dsl-db.pacinfo.com 1162229560 M * Bertl wb besonen_! 1162229592 M * sla 4307 rt_sigsuspend(~[HUP QUIT ALRM TERM TSTP RT_1] 1162229603 M * sla the only string appears 1162229658 M * Bertl hmm, rt is realtime .. strange ... 1162229710 M * sla ERROR 2002: Can't connect to local MySQL server through socket '/tmp/mysql.sock' (2 1162229717 Q * besonen Ping timeout: 480 seconds 1162229719 M * sla Check that mysqld is running and that the socket: '/tmp/mysql.sock' exists! 1162229778 M * Bertl and that exists and is fine (check with lsof) 1162230056 M * sla MMAMMAMMMAMMAAAAAAAAAAAAA MIA!!!!!!!!!!!! 1162230089 M * sla lsof produces 23MB output with over 30000 strings of mysql.sock 1162230182 M * Bertl interesting ... 1162230193 Q * stefani Quit: I Quit! 1162230195 M * doener yiihaa! 1162230287 M * sla ? 1162230323 M * ntrs sla, I had the same or similar problem once and it helped when I installed the latest mysql version downloaded from mysql.org, not the one that comes with the distribution. 1162230337 M * ntrs sla, which mysql version is that? 1162230461 M * sla I'm installing mysql from src. 4.0.27 1162230481 M * ntrs what distro? 1162230504 M * ntrs and do you see multiple mysql processes with ps aux or just a couple? 1162230578 M * ntrs If it's an rpm distro try installing the latest 4.1.x from rpm downloaded from mysql.org. That fixed the problem for a few of my guests. 1162230594 M * sla slack 1162230659 M * ntrs What I noticed was that if there were multiple processes in ps aux I had the problem. After I installed the rpm from mysql.org I only saw two mysql related processes, one of which was mysqld_safe and then there was no problem after that. 1162230681 M * ntrs I believe Bertl once said that it may have to do something with the type of threading used. 1162230684 M * doener ntrs: debian? 1162230689 M * ntrs nptl vs linux 1162230716 M * ntrs doener, no it was a rpm based distro, CentOS or Fedora 1162230724 M * ntrs I don't remember exactly 1162230725 M * sla why in the previous version of kernel/patch everything works? 1162230753 M * doener ok, there's a package similar to debian's libc6-i686 in these too I guess. (which add NPTL support) 1162230754 M * ntrs Could be a difference in the way threading works (nptl vs. linux threads) 1162230754 M * sla in any case, i'm compiling mysql from sources. 1162230778 M * sla and there are linux threads. 1162230788 M * ntrs sla, I am not sure but perhaps there is a compile option to choose one threading vs the other. 1162230856 Q * gerrit Quit: Client exiting 1162230916 M * sla threading... mysql on linux is working well, I'm using it from birthday. 1162230987 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162231061 M * sla Apache says: 4375 waitpid(-1, 0xbfd4c370, WNOHANG|WUNTRACED) = 0 1162231068 M * sla 4375 select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout) 1162231075 M * sla (via strace) 1162231270 M * Bertl sla: so you can run the _very same_ guest in (what version?) quite fine, while it barf on (other version)? 1162231427 M * sla hm.. I cannot understand... 1162231501 M * sla I kill apache, so, now I can mysql to mysqld. it shows, that mx conn. was 181 (of 500 possible) 1162231518 M * sla max conn. 1162231748 M * sla maybe, old glibc? 1162232062 Q * ruskie Ping timeout: 480 seconds 1162232226 J * coocoon ~coocoon@dslb-084-056-184-052.pools.arcor-ip.net 1162232250 M * coocoon hello to all 1162232355 Q * ntrs Remote host closed the connection 1162232361 M * doener Bertl: stacks grows down on x86, right? 1162232371 M * Bertl usually, yes 1162232374 M * doener ok 1162232756 J * hardwire ~hardwire@89-208-58-66.gci.net 1162233080 J * lilalinux ~plasma@dslb-084-058-200-181.pools.arcor-ip.net 1162233130 M * coocoon hello have a problem, i dunno exactly since when but I think ithas to do with latest util-vserver 30.211 version, on my guest there is shown that there are no space left on device buzt it isn't right as shown at this paste http://paste.linux-vserver.org/595 1162233138 M * coocoon shows 1162233146 Q * Piet Ping timeout: 480 seconds 1162233211 J * Piet hiddenserv@tor.noreply.org 1162233228 M * daniel_hozac coocoon: what does ls -ld /var/tmp show inside the guest? 1162233232 M * doener coocoon: does df on the host also say that there's space left? 1162233442 M * coocoon daniel_hozac: drwxrwxrwt 5 root root 4096 Oct 22 22:40 /var/tmp 1162233466 M * coocoon doener: http://paste.linux-vserver.org/596 1162233490 J * azuli1 ~Carlos@user-10cm30m.cable.mindspring.com 1162233670 M * coocoon daniel_hozac: is there anything wrong? 1162233769 M * Hollow doener, daniel_hozac, coocoon: simple: 1162233770 M * Hollow hera ~ # vdlimit --xid 500 /vservers/qkon/ 1162233770 M * Hollow 500 /vservers/qkon/ 1162233770 M * Hollow space_used=981532 1162233770 M * Hollow space_total=5242880 1162233771 M * Hollow inodes_used=100000 1162233772 M * Hollow inodes_total=100000 1162233787 M * Hollow coocoon: raise number of inodes... 1162233830 M * doener now I wonder where you got that output from 1162233838 M * Hollow i have access to his server ;) 1162233848 M * coocoon Hollow: thanx a lot 1162233851 M * doener cheater! 1162233882 M * Hollow well, it is a kind-of vcd test host 1162233901 M * Hollow (un)fortunately i have not used it for quite a time ;) 1162233972 Q * azuli1 Ping timeout: 480 seconds 1162234224 M * Loki|muh sorry for being offtopic, but maybe someone here knows a tool for drawing network plans? a describing language like latex would be great ;) 1162234297 M * Hollow i don't know a language, but maybe "dia" would help? 1162234363 J * bronson ~bronson@66.160.177.216 1162234674 J * stefani ~stefani@tsipoor.banerian.org 1162235256 J * ruskie ~ruskie@ruskie.user.oftc.net 1162235424 Q * sladen Ping timeout: 480 seconds 1162235661 J * sladen paul@starsky.19inch.net 1162235736 J * GoNoGo ~GoNoGo@82.236.93.215 1162235852 M * Bertl welcome GoNoGo! 1162235868 M * ensc Loki|muh: good old xfig... 1162235971 M * Loki|muh thanks, will try both 1162237539 Q * sla Quit: Leaving 1162238424 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1162239864 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1162241296 J * rgl Rui@217.129.151.190 1162241299 M * rgl hellloooo 1162241390 J * Hurga nobody@p508A84D8.dip0.t-ipconnect.de 1162241406 M * Hurga Hiya. 1162241609 J * Aiken ~james@tooax6-005.dialup.optusnet.com.au 1162241961 M * daniel_hozac hello 1162242053 J * DreamerC_ ~dreamerc@61-217-226-236.dynamic.hinet.net 1162242460 Q * DreamerC Ping timeout: 480 seconds 1162244324 M * rgl hi 1162244621 J * debugger Rui@217.129.151.190 1162245077 Q * rgl Ping timeout: 480 seconds 1162245279 Q * GoNoGo Quit: beam me up sco... 1162245477 M * micah I remember reading somewhere (maybe here on IRC?) about why someone would want to use vservers inside xen, does anyone recall this? 1162245507 M * micah or where the IRC logs are? 1162245555 M * Bertl irc.13thfloor.at 1162245599 M * mnemoc do you have any "favorite" irssi-to-html log processor? 1162245746 M * Bertl nope ,,, not yet 1162245933 M * mnemoc :\ 1162245985 M * Hurga micah: hang on 1162246036 M * derjohn Bertl, rc45 cool so far ? 1162246059 M * Bertl still some strange issues which are investigated right now 1162246101 M * Bertl the task lookup warnings showed lookups within the same guest as being blocked (which according to the code cannot happen, unless the task 'mutate') 1162246257 Q * ms_ Ping timeout: 480 seconds 1162246310 M * derjohn that was the mysql stuff, that was reported 1162246511 P * stefani I'm Parting (the water) 1162246587 Q * bonbons Quit: Leaving 1162246816 Q * bronson Quit: Ex-Chat 1162247095 Q * meandtheshell Quit: exit (0); 1162247874 M * Bertl derjohn: hmm? 1162247919 M * derjohn sla reported strange behavior on Slack with even rc45 1162247939 M * derjohn mysql and apache didnt behave as they did before 1162247963 M * Bertl well, we don't know what causes this, if that is 'caused' at all 1162247981 M * Bertl thing is, ntrs reported similar issues, which were fixed by a mysql upgrade 1162248000 M * Bertl note: I'm not saying it isn't an issue 1162248035 M * derjohn Because doener started to ask , i concluded that the mysql problem and the pid/process stuff is connected somehow. I might be totally wrong though ;) 1162248052 M * ntrs Bertl, it does not even have to be an upgrade I think, you can install the same version as the one you are running, only from mysql.org and not the distro delivered version 1162248081 M * ntrs It' seems they are compiled differently, nptl vs linux threads. 1162248105 M * ntrs This may have nothing to do with the problem really, it's just an observation 1162249222 Q * dna_ Quit: Verlassend 1162249490 J * bronson ~bronson@66.160.177.216 1162250843 Q * DreamerC_ Quit: leaving 1162250860 J * DreamerC ~dreamerc@61-217-226-236.dynamic.hinet.net 1162251408 Q * Hurga Remote host closed the connection 1162251501 Q * FireEgl Remote host closed the connection 1162251609 M * micah Hollow: ? 1162251620 M * micah err 1162251628 M * micah sorry