[00:25] mcp (~hightower@wolk-project.de) left irc: Ping timeout: 501 seconds [00:29] nathan_ (~nathan@209-6-130-26.c3-0.sbo-ubr1.sbo-ubr.ma.cable.rcn.com) left irc: Ping timeout: 480 seconds [01:21] ccooke (~ccooke@spc1-walt1-4-0-cust238.lond.broadband.ntl.com) joined #vserver. [01:21] ccooke_ (~ccooke@spc1-walt1-4-0-cust238.lond.broadband.ntl.com) left irc: Read error: Connection reset by peer [01:58] ozan (~ozan@dsl81-215-17504.adsl.ttnet.net.tr) joined #vserver. [02:00] hi ozan! [02:04] hi Bertl whats up? [02:04] i just try the 1.23 / 2.4.24 .. it crashes with no log [02:05] hmm, well, we are working on it, wan't to try something? [02:07] umm may be on my notebook, but cant try on servers ... i changed production server .. [02:08] please describe your server setup, where it crashed ... [02:08] - SMP or HT or UP ? [02:08] - do you use the 'shed' flag? [02:08] - do you have some monitoring tool querying /proc [02:09] ones is , Xeon(TM) CPU 2.40GHz [02:10] other is Pentium(R) 4 CPU 2.40GHz [02:10] both crashed? or better hung? [02:10] these two crashed with no msg , [02:10] i just use vtop ... [02:11] okay, after what uptime did they hang? [02:11] on xeon there is 10 vserver .. [02:11] umm let me check.. [02:13] xeonbox booted at 22:25 and 22:27 [02:14] p3 booted at 21:48 and crashed 22:25 [02:14] hmm that is pretty short .. [02:14] xeon box stands after first crash .. but i reboot it manually to rollback 1.1.6 [02:15] atm I would suggest going back to 1.22 but this will not fix the issue, it will only make it it little less likely to crash ... [02:17] umm ... it is better for me to stay with 1.1.6 :) these are production servers ... [02:17] well no problem with that ... [02:17] Nick change: surriel -> riel [02:18] by the way i found one of my server free, think i can test with that .. [02:18] it's simple that 1.1.6 is development, and 1.2x is stable branch, which means in stable are more bugs fixed than in the devel branch, doesn't say anything about the stability though ... [02:20] ozan: well if you want to test ... currently it seems that I can't reproduce the exact effect here (it seems very machine dependant) but some folks can ... and I'm currently waiting for somebody to help out ... [02:22] i want to help :) if i can ... i have installed that kernel on 5 server .. 2 of them crashed ... [02:23] but i dont wait others if they crash , i roll back .. [02:23] I have a tool to speedup the crash/hang ... [02:23] so if there is a free testing machine, which can be rebooted easily ... that would be optimal ... [02:24] we would have to test if this machine is prune for this kind of hang ... [02:24] and then try some modifications to the kernel (requires kernel rebuild) [02:25] umm this is a remote box, but tomorrow there will be 3 servers on my local .. may be we can do with them.. [02:27] when you online mostly? [02:29] well, depends, when I'm awake, I'm usually here ... [02:29] around 5pm - 6am CET [02:30] oki .. tomorrow i ll come, and we can test .. [02:30] good [02:31] i use vserver on all my servers :) it is vital for me [02:31] commercially? [02:32] umm i dont sell vserver but i use it on my hosting servers.. not for customers... [02:34] you know of the linux-vserver.org VServer Hosting page? did you already enter your company there? or under the VServer Users, or don't you like to add yourself there for whatever reason? [02:35] umm i just see , it no problem to be listed... [02:36] well it's no requirement, it's a feature, some kind of free promotion for your company and vserver ... [02:38] umm , it is nice to give vserver hosting .. this can be a new project for me .. [02:38] it can be a lot chep. [02:51] gotta go.. see ya tomorrow... [02:51] cu [02:51] so im back [02:51] good ... [02:51] but need to build in a new power supply to my devel. [02:52] :/ [02:52] the old one is up in smoke [02:52] ozan (~ozan@dsl81-215-17504.adsl.ttnet.net.tr) left irc: Quit: User abort with 5 Ctrl-C's [02:52] brb [02:52] to much killer ,) [02:52] hehe [03:57] ok maschine is up with a vs1.23 and 2.4.24er kernel (modded with the 2 lines) and killer still running for few minutes [04:01] you're still testing hard I see :) [04:06] sure :) [04:07] ps: killer is still running [04:07] okay, good ... we do another check ... [04:08] Sh[a]de: did you comment those lines out? [04:09] or just removed them? [04:09] 170 & 177 yes [04:09] comment out [04:12] oaky, please comment the in again, and change the lines between from: [04:12] if (p->s_info && (p->s_info->flags & VX_INFO_SCHED)) { [04:12] weight = atomic_read(&p->s_info->ticks) / [04:12] atomic_read(&p->s_info->refcount); [04:12] weight = (weight+p->counter) >> 1; [04:12] to [04:12] if (p->s_info && (p->s_info->flags & VX_INFO_SCHED)) { [04:12] weight = atomic_read(&p->s_info->ticks) / [04:13] (atomic_read(&p->s_info->refcount) + 1); [04:13] weight = (weight+p->counter) >> 1; [04:13] so what changes is the (atomic_read(&p->s_info->refcount) + 1); [04:14] ok done [04:15] if i can my eyes trust only the ") +1)" is added to this line, right?! [04:16] ok without the first ")" [04:16] ok done [04:16] hmm [04:16] still not right? [04:17] or why "hmm"? [04:17] *g [04:17] guess my version is correct ;) [04:17] weight = atomic_read(&p->s_info->ticks) / [04:17] (atomic_read(&p->s_info->refcount) + 1); [04:18] and there should be a '(' and a ' + 1)' ... [04:18] yea yea i see ok it's done [04:18] perfect ... ;) [04:19] compiler would complain anyway ... [04:19] and the locks are in place again, right? [04:20] you mean the "//" are deleted? [04:21] -yes [04:22] yes [04:22] these lines are again active [04:22] okay ... perfect [04:22] hey herbert [04:23] what wiki do you use on your servers? [04:23] tavi 0.25 + some changes ... [04:24] do you find it to be good? [04:27] ok running again with the modded kernel an killer running and crash after aprox 30sec [04:28] ctx: 58081 [04:28] and crash [04:29] it's pretty efficient ... [04:31] well... if you say this...i am calmed [04:31] hmm actually it seems my delay was extra large this time ... [04:31] because this was my answer to kestrel ;) [04:32] löl np [04:32] the thing is, I don't see how those locks could cause the hang ... [04:32] hmm [04:33] this error is only happend by me? [04:33] +is [04:33] well, it doesn't happen on my test scenarios ... [04:33] and I'm sure there is a good explanation for that ... [04:33] okay, suggestion: [04:34] put your .config file somewhere on the web, where I can download it [04:34] I'll compile a kernel for you and put it on my pages ... [04:34] the config from the kernel? [04:34] yes [04:34] ok thats np [04:37] need the current config? [04:37] i mean the config from kernel-2.4.24 with vs1.23? [04:37] yup, the one you use now ... [04:37] k [04:37] just one sec [04:37] filesystem check *gähn* [04:40] /arch/i386/config.in or? [04:42] or need the "defconfig" one? [04:43] it's .config in the main tree [04:43] directory [04:49] http://www.wwip.de/vs1.23-config [04:52] thanks ... [04:52] np [04:57] i go to take a shower ... [04:57] brb [05:44] Bertl i have to in bed [05:44] tomorrow we can testing again [05:45] for today i say thank u for spend time for me [05:45] okay, np ... [05:45] and sleep well [05:45] you too ... [05:45] ah [05:45] is the config right? [05:45] well, I used it and prepared a test kernel for you ... [05:46] ok [05:46] we test it tomorrow [05:46] sure, have a good night ... [05:46] now gn8 and sleep well all [05:46] thanks u too [05:46] <- away [06:40] Shade (shade@cpe109.bb101.cablesurf.de) joined #vserver. [06:47] Sh[a]de (shade@cpe109.bb101.cablesurf.de) left irc: Ping timeout: 492 seconds [07:43] Topic changed on #vserver by Bertl!~herbert@MAIL.13thfloor.at: http://linux-vserver.org/ || latest stable 1.24, devel 1.3.5, exp 0.04 [08:10] Shade (shade@cpe109.bb101.cablesurf.de) left irc: Killed (services.oftc.net (Too many invalid passwords)) [10:15] Nick change: Bertl -> Bertl_zZ [12:49] ensc (~ircensc@ultra.csn.tu-chemnitz.de) left irc: Ping timeout: 492 seconds [13:00] Auriga (~alanb@82-37-133-56.cable.ubr01.telf.blueyonder.co.uk) joined #vserver. [13:00] Good morning [14:10] noel (~noel@p50859C2B.dip.t-dialin.net) joined #vserver. [16:37] Doener_aw (~doener@pD9E129DB.dip.t-dialin.net) joined #vserver. [16:41] SURFKING (~XXXXXX@62.38.242.63) joined #vserver. [16:41] hello all [16:41] one small quick qs [16:42] is it possible to have 2 vserver and do fail over between them in a case of disaster? [16:44] Doener_zZz (~doener@pD9E12A70.dip.t-dialin.net) left irc: Ping timeout: 480 seconds [16:44] Nick change: Doener_aw -> Doener [16:48] quit [16:48] SURFKING (~XXXXXX@62.38.242.63) left irc: Quit: [x]chat [17:20] Shade (shade@cpe109.bb101.cablesurf.de) joined #vserver. [17:21] Nick change: Shade -> Sh[a]de [17:22] Auriga (~alanb@82-37-133-56.cable.ubr01.telf.blueyonder.co.uk) left irc: Quit: Good bye... Mister Anderson! [17:23] Sh[a]de (shade@cpe109.bb101.cablesurf.de) left irc: Quit: Excursion (On IRC.BONGSTER.DE [#wwip, #german-elite and #lov]) [17:33] Sh[a]de (shade@cpe109.bb101.cablesurf.de) joined #vserver. [17:33] re everybody [17:38] hmm [17:39] vproc seems not to work: readonly filesystem [17:39] but /proc on host is mounted rw [17:39] 1.24 [17:42] app-103:~/bin# ./vproc -d /proc/meminfo [17:42] ./vproc: read only filesystem. [17:42] proc on /proc type proc (rw) [17:43] hmm rmoriz, you should whait until Bertl or any of the devel. are online... my competence don't suffice for this :/ [17:44] i don't testet 1.24 for now... [17:55] click (click@gonnamakeyou.com) left irc: Ping timeout: 501 seconds [18:04] hm [18:04] isn't there a "reboot" thing on util-vserver? [18:04] from enrico? [18:05] hmm the "reboot" command is in vserver-029 from jack too, i think [18:13] click (click@gonnamakeyou.com) joined #vserver. [19:49] Bertl: I know u are still sleeping well ;) i want to say yout that we have found some guys who work now an a "PBVSC" it's a PHP Based V-Server Control and a website for this projekt already in progress... so will still working on it :) thats all and see ya later folks [20:41] Nick change: Bertl_zZ -> Bertl [20:42] hi all! [20:44] hi Bertl [20:45] was way too busy yesterday... could you send me the modifications you made to tavi? [20:45] yup, give me a second [20:45] sure [20:46] mail-to: bjoern.steinbrink@isp4p.net [20:47] hmm... writing too much html stuff... at least i didn't use a-tags ;) [20:47] serving (~serving@213.186.190.24) left irc: Read error: Connection reset by peer [20:48] serving (~serving@213.186.190.24) joined #vserver. [20:48] s/-// [20:48] Hi all [20:48] hi serving [20:48] hi! [20:48] :) [20:51] http://vserver.13thfloor.at/Stuff/patch-tavi-0.25-enhance3.diff [20:52] how is vs1.24? [20:58] Doener: not all changes are useful ... especially removing https isn't the 'final solution' ... [21:05] just got visitors... will be back shortly... [21:05] Nick change: Doener -> doener_na [21:07] nox (~nox@b109203.adsl.hansenet.de) left irc: Quit: nox [21:21] Here is -pre5. [21:21] This version fixes the "memory filled with unfreeable inodes" problem [21:21] which occurs on high memory machines in some workloads. That is the last [21:21] big problem 2.4 VM had with highmem I believe. [21:21] It contains a few other important fixes: [21:21] - fixes a SMP deadlock introduced during 2.4.23 (which could hang the [21:21] machine on high filesystem activity). [21:21] - fixes a memory allocation deadlock in USB [21:21] (from marcelo) [23:11] &w 23 [23:19] hmm, &w 23? [23:20] /w 23 [23:20] swap to window [23:21] ah [23:21] irssi [23:21] even if my ver-string says mirc [23:23] okay, have to leave now ... [23:23] cu later ... [23:23] later :) [23:24] damn [23:24] Nick change: Bertl -> Bertl_oO [23:24] :) [23:24] later Bertl_oO [23:25] rmoriz: : vproc works here on vs1.24 without any issues .-.- [23:25] are you sure you didn't mount the proc ro? [23:25] try with mount -o remount,rw /proc ... [23:26] cu later ... have to go now ... [23:38] Nick change: Bertl_oO -> Bertl [23:38] okay change in plans, I'm back ;) [23:39] rmoriz: what about your issues? [23:39] Nick change: doener_na -> Doener [23:41] hi Doener ! [23:41] hey Bertl [23:41] finally i have some time to start working on tavi :) [00:00] --- Sun Jan 18 2004