1266537603 M * Bertl e.g. full filesystem, journal=writeback mount and reboot
1266537653 M * Bushmills the disk with the zero size file was never full
1266537664 M * Bushmills had at least 200 gig still free
1266537695 M * Bushmills there's not just a few files zeroes. about all /bin /usr/bin /lib are zero
1266537725 M * Bushmills it was the mirror i was rsyncing to which run out of space
1266537756 M * Bushmills and the only reboot in recent days was the one for fscking the drive
1266537860 M * Bushmills http://stats.verhau.de/localdomain/localhost.localdomain/uptime.html
1266538061 M * Bushmills http://stats.verhau.de/localdomain/localhost.localdomain/diskstats_iops/sdb.html here is the rsync which caused the mirror disk to run out of space
1266538079 M * Bushmills and some examination following
1266538243 M * Bushmills here it shows the moment when sizes on first disk were suddenly not anymore what they were meant to be, manifestating by second disk running full:  http://stats.verhau.de/localdomain/localhost.localdomain/df.html
1266538267 M * Bertl as I said, without -S not really unexpected
1266538320 J * ViRUS ~mp@p579B5B3A.dip.t-dialin.net
1266538357 M * Bushmills well, ok. if that explains that all of the sudden second disk needs much more space for the same files, it wouldn't be related to apparently trashing .hash inodes on first disk
1266538452 M * Bertl try yourself, create a sparse file with e.g. dd if=/dev/zero bs=1M count=1 seek=1000000 of=/path/to/some/file
1266538485 M * Bertl then rsync it over, check the diskspace before and after
1266538514 M * Bushmills will that trash inodes on the first disk?
1266538521 M * Bushmills (source disk?)
1266538525 M * Bertl nope
1266538570 M * Bushmills i'll put the -S in. but i doubt that it is the cause of the problem i have with the source disk
1266538619 M * Bertl do you have disk limits in place?
1266538656 M * Bushmills no
1266538704 M * Bertl any messages in dmesg (from the time when files got trunkated)?
1266538720 M * Bushmills nope. i checked all logs.
1266538768 M * Bushmills but there was just getting no clue from those 
1266539232 M * daniel_hozac you do have --delete, right?
1266539810 M * Bushmills i just deleted the vserver.  
1266539830 M * Bushmills too messy to replace broken files. i better go the other way around. 
1266539861 M * Bushmills is  /etc/vservers/.defaults/vdirbase  important?   i don't have that one, but vserver guest delete  complained that it wasn't valid
1266539935 M * daniel_hozac i meant in your rsync
1266539935 M * Bushmills 2 of the 3 guests i had (the broken ones) were minimal systems anyway, no work involved there.
1266539949 M * Bushmills yes. actually i use that
1266539949 J * ViRUS_ ~mp@p579B5B5E.dip.t-dialin.net
1266540014 M * Bushmills rsync -aHvx --delete / /big/backup/sda    was my old invocation. follwoing Bertl's advice, I'll add -S to it 
1266540076 M * daniel_hozac vdirbase is important. it's how the utils will usually find your guests.
1266540084 M * daniel_hozac unless you specify the rootdir on every vserver ... build invocation.
1266540112 M * Bushmills vsroot? actually i do
1266540118 Q * yarihm Quit: Leaving
1266540120 Q * tpo Ping timeout: 480 seconds
1266540146 M * daniel_hozac yeah, you don't have to.
1266540147 M * Bushmills well, not by typing it in. but passed to newvserver in the script i use for that purpose
1266540188 M * Bushmills my vserver setup if relatively vanilla, pretty much out of the box
1266540203 M * daniel_hozac you shouldn't really use newvserver.
1266540255 M * Bertl well, if he likes to shoot himself in the foot :)
1266540255 M * Bushmills that's what the docs and descriptions i found suggested
1266540268 M * daniel_hozac where?
1266540275 M * Bushmills if there's something else and better, fine. 
1266540384 Q * ViRUS Ping timeout: 480 seconds
1266540403 M * Bushmills where .. i don't remember. i simply copied it into a script, substituted against command line args what i wanted to be able to specify, and kept it in my collection of scripts which deal with aspects of using vserver 
1266540484 M * Bushmills if that's not the right way, there were no red tape of warning signs
1266540487 M * daniel_hozac nothing in our docs should reference newvserver, nor Debian docs.
1266540509 M * daniel_hozac except maybe in a "it's deprecated" capacity.
1266540654 M * Bushmills why is it part of the package if it is deprecated?
1266540654 M * daniel_hozac it's not.
1266540654 M * daniel_hozac it's a separate package
1266540654 M * daniel_hozac the entirety of which is deprecated.
1266540686 M * Bushmills vserver-debiantools?
1266540689 M * daniel_hozac yes.
1266540726 M * Bushmills http://scarydevilmonastery.net/snap/1266540712440084363.png
1266540741 M * Bushmills if it says "deprecated", it doesn't do so very clearly
1266540773 M * daniel_hozac Ola doesn't think it is.
1266540790 M * daniel_hozac for whatever reason.
1266541031 M * Bushmills ah, i know why i didn't continue to use the wiki version, using vserver.   the meaning of  context wasn't clear to me.
1266541069 Q * dowdle Remote host closed the connection
1266541072 M * Bushmills i first thought that might me a bit mask, encoding a set of flags, thosen given a bit further down
1266541123 M * Bushmills when i found the newvserver version, which didn't need to specify such a mysterious number "42", i went along with that one
1266541204 M * Bushmills now that i have a better idea of what context represents, i think i should edit the wiki page in that respect
1266541228 M * Bushmills after all, that made me go the wrong route ...
1266541280 M * daniel_hozac you don't have to specify it. util-vserver will pick one for you if you don't.
1266541287 M * daniel_hozac which is what newvserver depends on.
1266541441 M * Bushmills i suppose the 42 was picked by a douglas adams fan, but it is not ideal in that context. it suggest that it needs to be a specific number. a number like 10000 appear to be less specific.  and i'd not be tempted to interpret 10000 as bit mask. 
1266541482 M * Bushmills but if it can be left out entirely, it should better be removed from the wiki "building a guest" command, as it only serves to confuse
1266542156 M * Bushmills good to know that vdirbase is needed .. between "installing" and "building a guest" on the wiki is no mention of it, but the building a guest command, as given there, fails because of vdirbase not exisiting.  had i not looked for another because of the "context", i has probably done at this point.
1266542256 M * Bushmills though vdirbase *is* mentioned in the FAQs .. in "How could I rename a vserver directory?"
1266542303 M * Bushmills so, what is it supposed to contain?
1266542306 M * Bertl 42 is the answer and it is as good a choice as any other number :)
1266542351 M * Bushmills i left it away from the guest building script now
1266542422 M * daniel_hozac vdirbase is, as the great flower page says, a symlink to where the guests live.
1266542430 M * Bushmills vdirbase is simply one line, containing the guest roots top level dir?
1266542432 M * daniel_hozac it is created on install.
1266542450 M * Bushmills ah. symlink.  sorry, not here. i don't have it.
1266542460 M * daniel_hozac then you didn't install the utils properly.
1266542523 M * Bushmills ah. my bad. other tools. i thought those where the obsolescent ones.
1266542779 M * Bushmills was installed. no idea why link wasn't there. collision with vserver-debiantools, maybe. but reinstalling helped, the link was there then.
1266543895 J * kraehe ~kraehe@77-21-172-61-dynip.superkabel.de
1266543952 M * kraehe moin I'm trying to setup a vserver on Gentoo - now on boot, it changes the hostname of my real system, when changing the hostname of the vserver - linux 2.6.32.8
1266543975 M * Bertl update util-vserver to something newer than two years :)
1266544022 M * Bertl i.e. switch to 0.30.216-pre instead of 0.30.215
1266544032 M * kraehe  0.30.215-r3 currently installed
1266544049 M * Bertl which is about 2 years old, and doesn't handle recent kernels
1266544057 M * kraehe *oh* ok
1266544077 M * Bertl also, you might want to mention that to the gentoo folks :)
1266544234 M * Bushmills with gentoo and debian both carrying obsolete packages .. what is the preferred vserver distrib at the moment?
1266544251 M * daniel_hozac Gentoo has recent versions as well.
1266544257 M * Bushmills masked?
1266544266 M * kraehe yes - just unmasked them - recompiling
1266544274 M * Bushmills ah, ok
1266544290 M * Bertl to answer the question, LVS :)
1266544299 M * Bushmills LFS?
1266544301 M * Bertl *LFS even
1266544332 M * Bertl any distro is fine, just compile the kernel and tools yourself and be happy
1266546854 M * kraehe *ok* the hostname part is fixed now - remaining problem is mysql "[ERROR] Can't start server: Bind on TCP/IP port: Address already in use"
1266547123 M * kraehe so - how to tell vserver not to share 127.0.0.1 between host and virtual?
1266547167 M * daniel_hozac it doesn't.
1266547174 M * daniel_hozac you didn't assign 127.0.0.1 to the guest, did you?
1266547217 M * kraehe i've installed the system by tutorial (rsync method from a xen system)
1266547262 M * Bertl single IP assigned to the guest?
1266547363 M * kraehe vserver morenga build -m rsync --context 42     --hostname morenga.kephra.de   --interface eth0:193.101.47.77/27        -- --source root@witboi:/
1266547429 M * Bertl your kernel probably has automatic single ip special casing turned on, so try putting ~single_ip in the guest configs nflags
1266547462 M * Bertl mysql tries to bind the IP _and_ 127.0.0.1, which end up being the same (with the single ip special casing on)
1266547512 M * kraehe where do I find those nflags?
1266547550 M * Bertl google for 'flower page' take the first result
1266547843 M * kraehe *ok* created a new file nflags containing ~single_ip - but still localhost is shared
1266547902 M * daniel_hozac did you restart the guest?
1266547907 M * kraehe yes
1266547944 M * Bertl where did you create that file?
1266547969 M * daniel_hozac what does nattribute --get --nid <guest> show?
1266547977 M * kraehe in /etc/vservers/morenga/nflags - and writing junk into that file, prevents starting the vserver - so i think i have the right place
1266548069 M * kraehe http://pastebin.com/m16394c71
1266548120 M * daniel_hozac and that is the only place you are running mysql?
1266548124 M * daniel_hozac not on the host?
1266548136 M * kraehe the host system is also running mysql
1266548141 M * daniel_hozac that'll be why.
1266548152 M * Bertl see FAQ page
1266548162 M * kraehe thats why its not starting on vserver - so: how to tell vserver not to share same localhost?
1266548175 M * Bertl that's not the problem
1266548179 M * daniel_hozac the host is bound everywhere.
1266548185 M * daniel_hozac on the guest's IP addresses.
1266548188 M * daniel_hozac you have to restrict it.
1266548191 M * kraehe ssh into morenga, and from there into localhost will ssh back to real host system
1266548199 M * Bertl the problem is that the host is bound to 0.0.0.0, including the guest ip 193.101.47.77
1266548274 M * Bertl same goes for sshd, needs to be restricted on the host too
1266548299 M * kraehe NO
1266548308 M * kraehe how to tell vserver not to share localhost
1266548329 M * daniel_hozac that's not your problem.
1266548332 M * daniel_hozac localhost is not your only IP.
1266548336 M * kraehe i want mysql on localhost, both on the real host and on the virtual systems
1266548384 M * daniel_hozac then you have to restrict the one running on the host to the host's IP.
1266548473 M * kraehe the other related question is routing/firewall - there is no backend (tun/tap) interface on host side created where the traffic of the virtual ethernet comes out
1266548496 M * kraehe or does vserver go out using the REAL ethernet?
1266548504 M * Bertl there is no 'virtual' ethernet in Linux-VServer 
1266548527 M * kraehe there are no virtual interfaces at all in vserver?
1266548531 M * Bertl Linux-VServer uses IP isolation, i.e. a subset of IP addresses is 'assigned' to the guest for 'use'
1266548562 M * Bertl you can get 'virtual' networking (and the overhead it brings) by using network namespaces
1266548568 M * kraehe but its not possible to assign same IP (e.g. 127.0.0.1) to several hosts?
1266548597 M * Bertl it is possible, but it will be shared, note that 127.x is handled special (lback remap and friends)
1266548637 M * Bertl i.e. a guest using lback remapping (default) will 'see' 127.0.0.1, but actually use a different IP instead (which is unique for each guest)
1266548694 M * kraehe so I can at least get distinct loopback with this lback remapping? how to enable this?
1266548721 M * Bertl it should be on by default, check with nattribute
1266548804 M * Bertl i.e. if your config stayed the same from the paste above, you have both lback_remap and hide_lback on
1266548813 M * kraehe its in nattribute - but how to use it?
1266548819 M * daniel_hozac it's already being used.
1266548885 M * kraehe we are turinging in cycles: how to tell vserver to share localhost is my question. So ho to tell it to use lback_remap to assign a differnt localhost to this vserver
1266548898 M * daniel_hozac it already has its own loopback.
1266548911 M * daniel_hozac as we said, that is not your problem.
1266548918 M * kraehe no - its shareing the loopback with the real host
1266548924 M * daniel_hozac the host is god.
1266548929 M * daniel_hozac it has all and sees all.
1266548936 M * daniel_hozac you can't do anything about that.
1266548954 M * daniel_hozac except tell your services to not mess with your guest's stuff.
1266548982 M * kraehe so its not possible to run several services on different servers on localhost
1266548988 M * daniel_hozac yes, it is.
1266549000 M * kraehe mysql always runs on localhost only - its normally not accessable from outside
1266549019 M * Bertl kraehe: please read what we write ... 
1266549042 M * Bertl the problem is that you have services running on your host, binding to _all_ IPs
1266549059 M * Bertl _all_ of course includes _all_ guest IPs
1266549084 M * kraehe all = all own IP (if I come form UML, or XEN, or OpenVZ point of view)
1266549096 M * Bertl normally you do not run any services on the host, services in a guest will be restricted to guest IPs
1266549116 M * Bertl even if they bind to 0.0.0.0, host services will _not_ be restricted
1266549158 M * Bertl both UML and XEN use a routed/switched setup, OVZ you can choose which one you want, but switched is default
1266549202 M * kraehe but vserver can not be used in a routed/switched/firewalled setup?
1266549223 M * daniel_hozac firewall has nothing to do with the first two.
1266549246 M * Bertl it can, as I wrote, you need to use network namespaces for that
1266549252 M * Bushmills ah,   dictd seems to be another program which is difficult to restrict, in the sense of binding it to more than one, but not all interfaces
1266549254 M * kraehe firewalling is done at both routing interfaces
1266549293 M * Bushmills to one interface is ok. to all is ok too. to, say,   loopback and eth0, but not eth1, is difficult 
1266549294 M * Bertl Bushmills, kraehe: putting such services into a guest solves that problem
1266549322 M * daniel_hozac Bushmills: primarily because that requires support for multiple sockets instead of one.
1266549333 M * Bertl (or even a network context does, if you want to keep it on the host)
1266549346 M * kraehe well - my idea was: I have a system at home, that normaly does not the vserver guests - those guests are backup from xen servers I have outside
1266549355 M * Bushmills yes. i just mention it because it wasn't listed in the wiki as problematic program.  bind was mentioned there, some others
1266549366 M * kraehe so my host is a full featured system
1266549386 M * Bertl Bushmills: so just add it and inform the maintain that it might need some more options :)
1266549397 M * Bushmills currently i just let it bind to all interfaces. unlikely that guest need their own instance
1266549407 M * Bushmills if they want one, there is one already :D
1266549484 M * Bushmills Bertl: right, though i might first look at alternative implemetations, which hopefully allow to bind to multiple but not all interfaces
1266549492 M * Bertl kraehe: you might be better off with a single xen/kvm domain running a Linux-VServer kernel then
1266549526 M * kraehe the host is an older CPU - so no kvm - i fear I have to use something like virtualbox
1266549533 M * Bushmills better to write "you can use this instead" rather than just "it won't work"
1266549533 M * kraehe or go back to user mode linux
1266549582 M * Bertl yes, UML is a good choice there too, you can put a Linux-VServer kernel into that too
1266549595 M * kraehe uml/skas was the way I used before - but I hoped that vserver has a better disc performance
1266549608 M * Bertl it definitely has
1266549639 M * Bertl as I said, you might also try to dig into network namespaces
1266549678 M * Bertl (not sure how far the util-vserver support goes there, didn't follow the testing)
1266549716 M * kraehe is there a plan to add routeable virtual interfaces to vserver?
1266549766 A * Bertl feels like going in circles ...
1266549774 A * kraehe still dreams about a virtual linux that can compete with solaris zones
1266549778 M * Bertl yeah, it's called network namespaces :)
1266549805 M * Bertl and guess what, it's already there :)
1266550347 M * kraehe but its hard to impossible to find information of how to use network namespaces with vserver
1266550405 M * Bertl the very same way without atm, i.e. I don't think the util-vserver integration is that far
1266550425 M * Bertl network namespaces are a mainline feature
1266550831 M * Bertl check it out, document it, report problems/issues/solutions here, blog about it, become famous :)
1266551782 J * petzsch1 ~markus@dslb-094-222-079-079.pools.arcor-ip.net
1266552069 J * SauLus_ ~SauLus@c192234.adsl.hansenet.de
1266552115 Q * petzsch Ping timeout: 480 seconds
1266552479 Q * SauLus Ping timeout: 480 seconds
1266552479 N * SauLus_ SauLus
1266555376 Q * ViRUS_ Quit: If there is Artificial Intelligence, then there's bound to be some artificial stupidity. (Thomas Edison)
1266558249 M * Bertl off to bed now .. have a good one everyone!
1266558256 N * Bertl Bertl_zZ
1266559900 J * ghislain ~AQUEOS@adsl2.aqueos.com
1266560312 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com
1266561393 Q * petzsch1 Quit: Leaving.
1266561873 M * incd Whats the best way to clone vserver container? vserver-copy seems to be outdated :)
1266561966 Q * Piet Remote host closed the connection
1266562032 J * Piet ~Piet__@04ZAAAISP.tor-irc.dnsbl.oftc.net
1266562491 J * ghislain1 ~AQUEOS@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr
1266562604 Q * ghislain Ping timeout: 480 seconds
1266562641 J * ncopa ~ncopa@245.39.189.109.customer.cdi.no
1266563632 M * BenG incd, here's one way: http://linux-vserver.org/Building_Guest_Systems#Building_guests_using_the_clone_build_method
1266566224 J * sharkjaw ~gab@90.149.121.45
1266568696 M * ghislain1 i had again a crash it seems 2.6.3x is not very stable. I think i will go back to 2.6.27 now
1266568763 M * pmjdebruijn 2.6.27 rocks :)
1266568807 M * ghislain1 yes it seems 2.6.3x has some big issues, time to go back in time i guess, we will see in 6 month if it is better....
1266568825 M * ghislain1 bertl: what 2.6.27 lacks over the last 2.6.32 patches ?
1266568837 M * ghislain1 bertl: i mean vserver wise of course
1266569261 M * pmjdebruijn no clue
1266569265 M * pmjdebruijn we're pretty content with 2.6.27
1266569296 J * tpo ~tpo@cable-dynamic-87-245-106-94.shinternet.ch
1266569682 M * BenG ghislain1, what are the circumstances of the crashes?
1266569709 M * BenG and are you getting the same with 2.6.31?
1266571114 Q * BenG Quit: I Leave
1266571221 J * thierryp ~thierry@lns-bzn-47f-62-147-212-202.adsl.proxad.net
1266571276 M * ghislain1 i have crashes in the network stack it seems
1266571304 M * ghislain1 seems the bug was introduced in 2.6.29 from what an audit said and is still not corrected
1266571320 M * ghislain1 let me check kernel version
1266571353 M * ghislain1  2.6.31.6-vs2.3.0.36.26 and 2.6.31.5-vs2.3.0.36.21
1266571386 M * ghislain1 both have complete  kernel freeze, seems related to network stack. They do not have the same nic one bnx other tg3
1266571454 J * barismetin ~barismeti@zanzibar.inria.fr
1266571855 J * bobnormal ~irc@87-194-32-179.bethere.co.uk
1266572015 J * dna ~dna@p54BC9CDE.dip0.t-ipconnect.de
1266572165 Q * tpo Ping timeout: 480 seconds
1266572998 P * kraehe Leaving
1266573843 Q * ncopa Ping timeout: 480 seconds
1266573988 Q * thierryp Remote host closed the connection
1266574318 J * ncopa ~ncopa@ti211310a081-1516.bb.online.no
1266574340 Q * balbir Read error: Connection reset by peer
1266574362 M * incd Hello again, I used this to clone vserver container: http://linux-vserver.org/Building_Guest_Systems#Building_guests_using_the_clone_build_method
1266574414 M * incd But it didn't do anything :p It just gave: vclone: "vc_set_iattr("/etc/vservers/.defaults/vdirbase/mail/dev/log"): No such file or directory"
1266574521 M * bobnormal incd: check that the /etc/vservers/.defaults/vdirbase symlink links to a real directory
1266574783 M * incd /etc/vservers/.defaults/vdirbase -> /vservers
1266574830 M * incd /etc/vservers/.defaults/vdirbase/ has only one vserver container. Its the one I'm trying to duplicate
1266574833 Q * Guest1917 Quit: ZNC by prozac - http://znc.sourceforge.net
1266574867 M * daniel_hozac what util-vserver version?
1266574875 J * Genghis ~Genghis@ph34r.my.d-n-s.org.uk
1266574896 N * Genghis Guest598
1266575199 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1266575220 J * gnuk_ ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1266575325 M * incd util-vserver: 0.30.215; Feb 12 2010, 14:26:23
1266575756 J * balbir ~balbir@122.172.49.152
1266576226 N * Bertl_zZ Bertl
1266576231 M * Bertl morning folks!
1266576253 M * Bertl incd: too old for kernels newer than 2.6.22
1266576923 M * incd ok, thanks again
1266577071 M * incd util-vserver-0.30.216-pre2864 then?
1266577297 M * Bertl yep, that should be fine with recent kernels
1266577751 M * incd Oki
1266578191 Q * bobnormal Ping timeout: 480 seconds
1266579383 Q * weasel Read error: Connection reset by peer
1266579395 J * weasel ~weasel@weasel.noc.oftc.net
1266580300 Q * Piet Ping timeout: 480 seconds
1266582402 M * ghislain1 bertl: hello :) is there any functional differences between latest vserver 2.6.32 and latest 2.6.27 ?
1266582454 M * Bertl yes, 2.6.32 has a bunch of features not available on 2.6.27
1266582666 J * petzsch ~markus@dslb-094-222-079-079.pools.arcor-ip.net
1266583290 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com
1266583638 M * ghislain1 bertl: i saw cgroup meme limits are now used to show memory available this is cool
1266583653 M * ghislain1 bertl: i fear using again a 2.6.32 :(
1266583744 M * petzsch always did, sorry for that false info in the wiki, when i tested i set the limits live in /dev/ and that didn't seem to show up in top
1266583764 M * petzsch i'll remove that line from the wiki, sorry for the confusion
1266583766 M * Bertl ghislain1: well, 2.6.31 better?
1266583781 M * petzsch no problems with 2.6.32 here
1266583885 M * ghislain1  petzsch: do you use it on dell machines ?
1266583973 M * petzsch nope, on a Hetzner EQ4 root server
1266583988 M * ghislain1 seems related to network stack happen in random situation , ah ok
1266584051 M * petzsch has anyone experience with using vhashify in an openvcp envoirment? does their backup handling breakt anything i should be aware of?
1266584296 M * BenG ghislain1, what are the 2.6.32 problems of which you speak?
1266584399 M * ghislain1 Beng: random freeze of the kernel from what we can see on the screen it *SEEMS* related to a bug in network stack introduced by 2.6.29 (from an audit we made)
1266584419 M * Bertl audit means?
1266584481 M * ghislain1 bertl: mean the autors of haproxy were called to help on the issue looked at it and said that
1266584511 M * ghislain1 bertl: at first we thinked it was an haproxy issue
1266584532 M * Bertl i.c. so it was narrowed down to a bunch of patches then?
1266584551 M * ghislain1 they linked us to this kernel bug: http://bugzilla.kernel.org/show_bug.cgi?id=14470
1266584663 M * ghislain1 this is mainline not vserver of course
1266584914 M * ard petzsch : those hetzners, are they supermicros?
1266584964 J * bobnormal ~irc@87-194-70-98.bethere.co.uk
1266584975 M * BenG looks like may people are getting that same bug
1266584993 M * BenG s/may/many/
1266585013 M * BenG I've not seen anything like that
1266585027 A * ard also
1266585033 M * ard so I am glad for the heads up
1266585034 M * ard :-)
1266585047 M * BenG absolutley
1266585059 M * ard Our high traffic servers do about 1.5 gig per server
1266585090 M * BenG in what time period?
1266585099 M * ard bits/second
1266585115 M * petzsch according to dmidecode hetzner eq4 is based on a MSI X58 Pro-E board
1266585124 M * BenG and which kernel are you on ard?
1266585141 M * ard those high traffic certainly are not 2.6.32 :-)
1266585156 M * ard more like 2.6.26 or so with netpatches merged from dave miller
1266585174 A * ard checks to see
1266585254 M * ghislain1 i have several on 2.6.31 and 2.6.32 that works ok but this is the third box to have the issue now as time goes by more machines are affected
1266585293 M * BenG so some are crashing more often than others ghislain1?
1266585475 M * ard I do got 2 servers which do 100Mb/s...900Mb/s under memory stress conditions. 2.6.32.3 with vserver patches. and they seem stable
1266585507 M * ghislain1 yes
1266585588 M * ghislain1 as soon as i upgraded 2 boxes to 2.6.31 they started crashing, now i have a third box doing the same on 2.6.32. Old kernel woked fine for month
1266585608 M * ghislain1 so i will have to revert to old kernel :)
1266585642 M * BenG server here 2.6.31.5-vs2.3.0.36.22 with an uptime of 102 days
1266585670 M * BenG and another with 101 days
1266585718 M * BenG ghislain1, are any of those that crash running the kernels I compiled that live on repo.psand.net
1266585720 M * BenG ?
1266585728 M * Bertl ghislain1: did you test with a different nic yet?
1266585792 M * ghislain1 no i had not the oportunity to do that 
1266585809 M * Bertl IMHO that would be most interesting
1266585835 M * Bertl shouldn't be too hard to get e.g. a tg3 or similar instead of the intel nic (or in addition?)
1266585859 M * ard e1001e? There was a bug fixed for the e1001e in 2.6.32.8 or 7
1266585861 M * ghislain1 beng: no this is still my own  kernels not yours
1266585883 M * ard my servers have bnx2 except for the high-traffice (because bnx2 cannot handle high traffic)
1266585954 M * BenG NetXtreme II BCM5708, therefore bnx2 here also
1266587049 M * ghislain1 ok thanks for your input, i cannot change all the nic i have on my dell and on customers dell machines but i will try to test this as soon as i can :)
1266587118 M * Bertl change it on a machine which crashes often
1266587137 M * Bertl it might allow to verify that it is a nic/driver issue
1266587169 M * ghislain1 yes i have ordered some other nic for that i have an hardware maintenance the 1st march so i will try that at this moment :)
1266589061 J * thierryp ~thierry@lns-bzn-47f-62-147-212-202.adsl.proxad.net
1266589252 Q * BenG Quit: I Leave
1266589795 Q * hijacker Quit: Leaving
1266589801 Q * sharkjaw Remote host closed the connection
1266590267 J * tpo ~tpo@cable-dynamic-87-245-106-94.shinternet.ch
1266591443 Q * tpo Ping timeout: 480 seconds
1266591711 J * Piet ~Piet__@04ZAAAI6D.tor-irc.dnsbl.oftc.net
1266591903 Q * thierryp Ping timeout: 480 seconds
1266594469 Q * kjj Quit: arf
1266595295 Q * AndrewLee Read error: Connection reset by peer
1266595473 J * AndrewLee ~andrew@u7.hlc.edu.tw
1266595476 Q * AndrewLee 
1266595480 J * AndrewLee ~andrew@u7.hlc.edu.tw
1266595528 Q * AndrewLee 
1266595537 J * AndrewLee ~andrew@u7.hlc.edu.tw
1266595778 J * dowdle ~dowdle@scott.coe.montana.edu
1266597810 J * derjohn_mob aj@80.187.150.47
1266597811 Q * derjohn_mob 
1266597824 J * derjohn_mob aj@80.187.150.47
1266598126 J * imcsk8 ~ichavero@148.229.1.11
1266598163 J * ghislain ~AQUEOS@adsl2.aqueos.com
1266598468 Q * ghislain1 Ping timeout: 480 seconds
1266599046 M * bobnormal very happy, just got vserver host running nginx dispatching proxied http request through to 0.0.0.0-bound vservers (ie: portable / lacking in local config) to server load-balanced pages with failover support
1266599256 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1266599571 M * theocrite bobnormal: why wouldn't you install your nginx in a guest?
1266599575 A * theocrite does that
1266599610 J * thierryp ~thierry@home.parmentelat.net
1266599696 Q * bonbons Quit: Leaving
1266599861 Q * thierryp Remote host closed the connection
1266600047 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1266600540 Q * ncopa Quit: Ex-Chat
1266600573 M * bobnormal theocrite: could do, maybe i will :)
1266601145 Q * barismetin Quit: Leaving...
1266601381 J * thierryp ~thierry@home.parmentelat.net
1266602308 Q * bobnormal Quit: verymerry weekend!
1266603795 Q * thierryp Remote host closed the connection
1266603812 Q * imcsk8 Quit: Leaving
1266604156 Q * gnuk Quit: NoFeature
1266604158 Q * gnuk_ Quit: NoFeature
1266604610 Q * derjohn_mob Ping timeout: 480 seconds
1266605674 J * hijacker ~hijacker@87-126-142-51.btc-net.bg
1266605907 J * thierryp ~thierry@home.parmentelat.net
1266606088 Q * dna Quit: Verlassend
1266607047 Q * bonbons Quit: Leaving
1266607178 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1266607657 Q * thierryp Remote host closed the connection
1266608199 Q * FireEgl Quit: Leaving...
1266610022 Q * Marillion Remote host closed the connection
1266610024 J * Marillion ~dirk@hetzner4.127011.net
1266610036 Q * [Guy] Remote host closed the connection
1266610042 J * Guy- ~korn@elan.rulez.org
1266610055 Q * biz Remote host closed the connection
1266610058 J * biz ~biz@node2.cluster1.pyrox.eu
1266610230 J * derjohn_mob ~aj@e180195128.adsl.alicedsl.de
1266610422 M * Bertl off to bed now ... have a good one everyone!
1266610427 N * Bertl Bertl_zZ
1266611236 J * thierryp ~thierry@home.parmentelat.net
1266611305 Q * thierryp 
1266611741 J * ktwilight ~keliew@177.58-240-81.adsl-dyn.isp.belgacom.be
1266611954 Q * ktwilight__ Ping timeout: 480 seconds
1266612002 J * krushik ~krushik@modny-spb.ru
1266612233 Q * SauLus Ping timeout: 480 seconds
1266612817 Q * krushik Quit: Lost terminal
1266613570 J * tpo ~tpo@cable-dynamic-87-245-106-94.shinternet.ch
1266613644 J * cehteh` ~ct@pipapo.org
1266614205 Q * cehteh Quit: Terminated with extreme prejudice - dircproxy 1.0.5
1266614279 N * cehteh` cehteh
1266617439 J * SauLus ~SauLus@c152047.adsl.hansenet.de
1266617721 Q * hijacker Quit: Leaving
1266617882 J * SauLus_ ~SauLus@c152043.adsl.hansenet.de
1266618290 Q * SauLus Ping timeout: 480 seconds
1266618290 N * SauLus_ SauLus
1266618375 J * SauLus_ ~SauLus@c207035.adsl.hansenet.de
1266618775 Q * SauLus Ping timeout: 480 seconds
1266618775 N * SauLus_ SauLus
1266618840 J * SauLus_ ~SauLus@d004154.adsl.hansenet.de
1266619061 Q * bonbons Quit: Leaving
1266619260 Q * SauLus Ping timeout: 480 seconds
1266619260 N * SauLus_ SauLus
1266619432 J * krushik ~krushik@modny-spb.ru
1266620092 J * SauLus_ ~SauLus@c192105.adsl.hansenet.de
1266620318 M * urbee http://linux-vserver.org/Getting_high_with_lenny
1266620320 M * urbee uhh
1266620321 M * urbee i like!
1266620480 J * dna ~dna@p54BC9CDE.dip0.t-ipconnect.de
1266620500 Q * SauLus Ping timeout: 480 seconds
1266620500 N * SauLus_ SauLus
1266621880 Q * dna Quit: Verlassend
1266623202 J * rickytato ~rickytato@host100-164-dynamic.41-79-r.retail.telecomitalia.it
1266623544 J * SauLus_ ~SauLus@c192187.adsl.hansenet.de
1266623955 Q * SauLus Ping timeout: 480 seconds
1266623955 N * SauLus_ SauLus