[00:09] are you in a position to debug that problem now? It might be something superficial [00:10] Sure... [00:10] I will post my config... [00:10] S_CONTEXT=10 [00:10] IPROOT="eth0(eth0):1.2.3.4/255.255.255.128" [00:10] ONBOOT=yes [00:11] S_HOSTNAME=host.domain.name [00:11] S_DOMAINNAME=none [00:11] #S_NICE= [00:11] S_FLAGS="lock nproc sched" [00:11] ULIMIT="-H -u 150" [00:11] S_CAPS="CAP_NET_RAW CAP_NET_ADMIN" [00:11] That's it.. [00:11] Obviously the hostname and the addresses are real ones.. [00:12] what error do you get? [00:12] I had it working on another machine, but blew it away.. Didn't think to save the config.. [00:12] New security context is 10 [00:12] Setting network parameters: [ OK ] [00:12] Bringing up loopback interface: SIOCSIFFLAGS: Permission denied [00:12] Failed to bring up lo. [00:12] [FAILED] [00:12] Bringing up interface eth0: SIOCSIFFLAGS: Permission denied [00:12] Failed to bring up eth0. [00:12] [FAILED] [00:12] That is what I get from the host side when I start the vserver. [00:16] That's when you start a vserver? Does `chbind' work? [00:17] what about `ifconfig eth0:foo 1.2.3.4 up' ? [00:23] No, I am using Shadow (Alexey's) version, which is a bit different... [00:23] In his tree, the virtual interface is not viewable from the host... [00:23] In addition it doesn't us chbind.. [00:30] surely the kernel that does work on the machine has a config available for it, especially if it's a standard RH kernel [00:34] Action: mugwump & # switching contexts [00:54] shuri (~ipv6@CroCrodile.HuNter.blacktaboovideo.com) left irc: Remote host closed the connection [01:58] mlo (~ml@80.136.179.180) left irc: Quit: Leaving [02:36] rjg (~rjg@68.66.145.169) left irc: Remote host closed the connection [02:37] ChuckD_zZ (~bug@CPE-144-137-113-118.nsw.bigpond.net.au) left irc: [03:19] shuri (~ipv6@3ffe:bc0:8000::5bf) joined #vserver. [03:19] hello [03:20] usr/local/sbin/vserver: ulimit: connot modify man user process: Invalid argument [03:20] with v-1.0 [03:48] ShuX (~funny@207.236.226.187) joined #vserver. [03:54] ShuX (~funny@207.236.226.187) left irc: Read error: Connection reset by peer [04:17] shuri: change -H to -SH in the conf file [04:28] serving (~serving@213.186.190.110) left irc: Ping timeout: 485 seconds [05:00] serving (~serving@213.186.190.24) joined #vserver. [05:54] shuri (~ipv6@3ffe:bc0:8000::5bf) left irc: Remote host closed the connection [06:46] Nick change: Bertl_oO -> Bertl_zZ [07:03] Medivh (ck@server1.shell-express.de) left irc: Ping timeout: 492 seconds [07:11] Medivh (ck@195.167.218.4) joined #vserver. [07:24] Medivh (ck@195.167.218.4) left irc: Ping timeout: 485 seconds [07:25] Medivh (ck@server1.shell-express.de) joined #vserver. [08:27] Medivh (ck@server1.shell-express.de) left irc: Quit: changing servers [08:39] Medivh (~ck@195.167.218.4) joined #vserver. [08:42] Medivh (~ck@195.167.218.4) left irc: Quit: changing servers [08:54] Medivh (ck@server1.shell-express.de) joined #vserver. [08:56] rjg (~rjg@fl-atlnfl-u2-c3a-169.atlsfl.adelphia.net) joined #vserver. [09:15] rjg (~rjg@fl-atlnfl-u2-c3a-169.atlsfl.adelphia.net) left irc: Remote host closed the connection [09:37] memzy (desolation@68.154.36.99) joined #vserver. [09:37] hi hi [11:38] memzy (desolation@68.154.36.99) left irc: Quit: Nap time for this kid! Night y'all take good care! [12:35] kestrel_ (~athomas@dialup28.optus.net.au) joined #vserver. [12:49] Nick change: dst_ -> dst [12:50] Nick change: dst -> knurl [12:50] knurl (~dst@pD9E3995B.dip.t-dialin.net) left irc: Quit: leaving [18:06] Nick change: Bertl_zZ -> Bertl [18:06] hi all! [18:20] hi [18:24] hi matt! [18:33] gaertner (~gaertner@212.68.83.129) left irc: Remote host closed the connection [18:34] hello [18:34] any news regarding the 'critical' server? [18:35] it's up [18:35] i expect in a few days kswapd will take a shit [18:35] neither nesh or mrbawb have tested my patchset on SMP... but I sent it to both [18:35] hmm, so the O(1)/rmap testing is on hold then ... [18:37] yeah [18:38] but you still have it runnung on a UP machine? [18:38] that server is running 2.4.22-c17e though, running great with 1600-1900 procs and 50 vservers [18:38] i imagine once O(1)+rmap work on it it'll be even better [18:38] eyah i'm running 2.4.23-pre8+c17h(O1+rmap+ml+dl+vr) and it's great [18:39] 550MB swap in use, .23 average load [18:39] ayn change in the patches you used there (besides SMP changes) [18:39] nope [18:40] you mean between the 2 servers? [18:40] yes ... [18:40] it's the same exact patches... i have a patch directory with a perl script to patch in a certain order [18:40] hmm, and it's unlikely that you updated one of those patches, right? [18:40] i didn't [18:41] their the same :) [18:41] okay, just wanted to know after what we are going ... [18:41] the only problem is I still have to include major.h in quota.c [18:42] manually... i didn't make a patch to "fix" that [18:42] that will go away when the vr is released as stable ;) [18:45] i'd just like to see if another SMP server crashes with my patches [18:45] harder to find someone with SMP willing to test than I thought :) [18:46] Whom are you telling this ;) [18:47] but soon I'll have a test system (SMP) up and running ... [18:48] yeah? [18:49] hmm, why do you need to include major.h in quota.c (now that I think of it?) [18:50] ahh okay, forget it ... found it ;) [18:52] what would you consider a useful vroot devicenumber (for default)? [18:53] rephrased: how many vroot devices would you like to have if nothing is specified? [18:54] hrm [18:54] would this be a kernel config option? [18:54] i'd say 16-32 is a good default, personally i'd just max it out and i'm unsure [18:54] ideally i'd like 128 per server :) [18:55] er, as i'm unsure [18:55] hmm, I could make this a config option, but if I look at the loop device, there is no such option for example ... [18:56] the default there is 8 and it can be overridden on the boot command line ... [18:56] or specified on insmod, if compiled as module [18:59] oh yeah [18:59] well, as long as it can be set to a fairly high number [18:59] i'd aim for 256 [18:59] or more [18:59] never know if someone is running vserver on a quad xeon with 16GB of ram [18:59] with all unified vservers [19:00] 256 is the current maximum ... [19:00] they could prolly run a thousand vservers... [19:00] and two unified vservers should use the same vroot ... [19:00] hrm [19:00] ok, then all NOT unified :) [19:00] oh yes [19:00] probably not easy to get the 256 LVM partitions working ;) [19:01] 1 vroot per block device [19:01] in that case i'd max out at needing 4 :) [19:01] forgot about that.. [19:01] okay, I set it to 8 (like the loop device) and keep the override .. ;) [19:02] increadible ... 9/11/2002 (this is when vroot was written ;) [19:03] hahahaha [19:03] hadn't thought that it is almost a year, since I wrote that code ... [19:14] time flies when you're having fun ;) [19:15] and it won't move, if you are bored 8-) [19:17] @riel your 2.6 vserver efford was canceled, right? [19:17] or just delayed to 2005? [19:17] for now, yes ;( [19:17] delayed I suspect [19:17] unless I run into another very efficient virtualisation project [19:18] okay, so I don't step on your toes if I port some pieces ... [19:18] nope [19:18] want my edited selinux hooks.c for vserver ? ;) [19:18] sure ... [19:20] on the positive side, I will be focussing on CKRM now [19:20] so when your 2.6 vserver stuff is done, there might be a resource management framework to combine it with [19:20] would be nice ... [19:21] hope you will stay around on the channel, just to share/discuss some ideas ... [19:21] I will [19:21] what's your email address again ? [19:21] herbert@13thfloor.at [19:22] riel: if you mail it to vserver@paul.sladen.org I'll stick it in the contrib directory [19:22] netrose (~john877@24.171.21.47) left irc: Ping timeout: 485 seconds [19:22] sladen: it doesn't even compile yet I'm afraid [19:22] but it's partly there [19:23] I'll mail it to you anyway, but just for you to look at [19:23] hey paul, what is sitting in your contrib dir? [19:23] better not put it online for now [19:24] sent [19:24] tx [19:24] @sladen could you give some comments to this stuff? [19:25] @sladen You don't have permission to access /vserver/contrib/ctx16-sched-BUG-fix/ on this server. ? [19:29] bertl: fixed [19:30] bertl: it's mostly stuff I've picked up off the mailing list and want to be sure doesn't get completely list [19:30] s/list/lost/ [19:30] yeah I'm reading through it .. and I'm surprised how many ideas are there ... [19:33] ccooke (~ccooke@80.1.164.238) left irc: Ping timeout: 485 seconds [19:37] riel: ta [19:38] np [19:49] sladen: what happened to the sys_reboot stuff? [19:49] oh, uh a level [19:49] er, up [20:02] matta: I have no spare hardware for testing kernels on--it's code that I wrote, compiled; but have never had the opportunity to test [20:04] i'll have to test that sometime.. lots of people ask why reboot doesn't work for them :) [20:04] hmm, so a virtualized reboot would be a GoodThing(tm)? [20:05] bertl: yes. Because then rebootmgr, and the crappy vreboot/vhalt disappear [20:05] I'm not sure that the rebootmanager would disappear, but I could think of something like the hotplug agent ... [20:06] read the code!!! (I wrote it over a year ago...) [20:06] http://www.paul.sladen.org/vserver/sys_reboot/ [20:06] I'll do that ... [20:06] yeah, i don't like the whole rebootmgr process either [20:06] i actually wrote a web interface for users to restart their servers [20:07] i still run rebootmgr, but i don't really support it's use [20:07] also, since rebootmgr is started with the vservers on stdin [20:07] it's limited [20:07] what is stdin's limit? 2048 chars? [20:08] so once you have many vservers it will just stop working for the ones at the end of the alphabet [20:08] ctx >= 0 && uid == 0 | Real reboot | Fake reboot [20:08] is this something we want? real reboot if CAP_SYS_BOOT? [20:08] i personally do not [20:09] well, i guess it makes sense that if you allow it it should be able too.. [20:10] hmm, looks like what I had in mind ... will code up a patch for the devel tree ... [20:15] Action: Bertl is reading fix0r.txt ... [20:16] @sladen private filesystem loopback devices: this is a security issue, not a technical one ... [20:28] AGoe (~agoeres@Deb17.d.pppool.de) joined #vserver. [20:29] Hello everybody, Hi Herbert [20:31] AGoe (~agoeres@Deb17.d.pppool.de) left irc: Client Quit [20:40] hmm, this must be some hobby, short time appearances ... [21:17] netrose (~john877@24.171.21.47) joined #vserver. [21:25] okay cu l8er ... [21:25] Nick change: Bertl -> Bertl_oO [21:29] gaertner (~gaertner@212.68.83.129) joined #vserver. [22:51] netrose (~john877@24.171.21.47) left irc: Ping timeout: 485 seconds [23:32] Whee! 1:17 to compile a 2.4.22-ac4-c17g4 kernel on the dual Opteron B-) [23:49] kestrel_ (~athomas@dialup28.optus.net.au) left irc: Quit: blaha [00:00] --- Mon Nov 3 2003