[00:01] JonB (~NoSuchUse@129.142.112.33.ip.tele2adsl.dk) left irc: Ping timeout: 488 seconds [00:09] talon: [00:09] # /bin/mount --bind -o ro /mnt/part1 /mnt/part3 [00:09] # touch /mnt/part3/test/existing_file [00:09] touch: /mnt/part3/test/existing_file: Read-only file system [00:09] # /bin/umount /mnt/part3 [00:09] does this work for you? [00:11] root@test1:~# cd /root [00:11] root@test1:~# touch 1/foo [00:11] root@test1:~# mount -oro --bind 1 2 [00:11] root@test1:~# touch 2/foo [00:11] touch: cannot touch `2/foo': Read-only file system [00:11] root@test1:~# umount 2 [00:11] umount: /root/2: device is busy [00:12] Last message repeated 1 time(s). [00:12] hmm, interesting, works with 2.4.23-bme0.03 here ... [00:13] are part1 and part3 differetn filesystems ? [00:13] mount -t ext2 -o rw /dev/discs/disc1/part1 /mnt/part1 [00:13] /mnt is on / [00:13] ok let me try it with an ext2 fs. [00:13] im just doing this on root which is ext3 [00:14] hmm, thing is I can reproduce, if I ahve the whole test running on that mount, so it _is_ an issue, but I would like to narrow it down somewhat ... [00:18] root@test1:~# mkdir /mnt/part1 /mnt/part2 [00:18] root@test1:~# mount -t ext2 -orw /dev/sdb1 /mnt/part1 [00:18] root@test1:~# cd /mnt [00:18] root@test1:/mnt# touch part1/foo [00:18] root@test1:/mnt# mount -oro --bind /mnt/part1 /mnt/part2 [00:18] root@test1:/mnt# touch /mnt/part2/foo [00:18] touch: cannot touch `/mnt/part2/foo': Read-only file system [00:18] root@test1:/mnt# umount /mnt/part2 [00:18] umount: /mnt/part2: device is busy [00:18] Last message repeated 1 time(s). [00:18] which kernel? [00:19] 2.4.23 [00:19] hmm, interesting ... [00:19] i can try it again with the 2.4.23 without quotas enabled. [00:19] you want my kernel config file ? [00:19] what about that theory: you touched a file _before_ the bind mount was done, and it is still in inode cache? [00:20] now you access that file/dir through the ro --bind mount ... [00:21] sounds about right except fro thse case where i create dirs from scratch and do ro mount and touch a non existing file. [00:22] i really dont know much about teh kernel internals involved. [00:23] no i guess that would touch the cache too. [00:24] but i have mounted after a reboot those same empty dirs as well. [00:24] and cause it to go busy. [00:33] after reboot. [00:33] root@test1:~# mount -t ext2 -orw /dev/sdb1 /mnt/part1 [00:33] root@test1:~# mount -oro --bind /mnt/part1 /mnt/part2 [00:33] root@test1:~# touch /mnt/part2/foo [00:33] touch: cannot touch `/mnt/part2/foo': Read-only file system [00:33] root@test1:~# umount /mnt/part2 [00:33] umount: /mnt/part2: device is busy [00:33] Last message repeated 1 time(s). [00:33] with foo being there form last time. [00:34] erm wait a second. [00:34] foo not being there would be better ;) [00:36] doh... ok foo was always there. becuase its the vservers fs. [00:36] and foo is the name of the ctx100 vserver... [00:36] hmm a directory? [00:36] only ext2fs i have. [00:36] yeah. [00:36] nothing else I test, yoda ;) [00:37] # mount -t ext2 -o rw /dev/discs/disc1/part1 /mnt/part1 [00:37] # /bin/mount --bind -o ro /mnt/part1 /mnt/part3 [00:37] # touch /mnt/part3/test/non_existing_file [00:37] touch: /mnt/part3/test/non_existing_file: Read-only file system [00:37] # /bin/umount /mnt/part3 [00:37] umount: /mnt/part3: Device or resource busy [00:38] so it happens when I touch a file through ro --bind which doesn't exist ... [00:49] root@test1:~# mount -t ext2 -orw /dev/sdb1 /mnt/part1 [00:49] root@test1:~# mount -oro --bind /mnt/part1 /mnt/part2 [00:49] root@test1:~# touch /mnt/part2/foo/quota.txt [00:49] touch: cannot touch `/mnt/part2/foo/quota.txt': Read-only file system [00:49] root@test1:~# umount /mnt/part2 [00:49] umount: /mnt/part2: device is busy [00:49] Last message repeated 1 time(s). [00:49] root@test1:~# ls /mnt/part2/foo/quota.txt [00:49] /mnt/part2/foo/quota.txt [00:50] so teh existign file case doesnt seem to work for me. [01:02] WSU (~Josh@ny.webpipe.net) left irc: Quit: Leaving [01:02] although im guessing your using lvm. instead of raw partitions. [01:03] no idea if htat makes any difference. [01:03] nope just raw partitions ;) [01:03] naming is only because I use devfs ;) [01:08] dont even think i have devfs suport compiled in. [01:08] that should not make any difference, the block device is the same ;) [01:09] dont think it makes any difference either. [01:09] JonB (~NoSuchUse@129.142.112.33.ip.tele2adsl.dk) joined #vserver. [01:09] hi Jon! [01:09] hey Bertl [01:18] serving (~serving@213.186.188.205) joined #vserver. [01:20] arekm (misiek@ikar.t17.ds.pwr.wroc.pl) joined #vserver. [01:20] arekm_ (misiek@ikar.t17.ds.pwr.wroc.pl) left irc: Read error: Connection reset by peer [01:28] kestrel (athomas@home.swapoff.org) left irc: Quit: ircII EPIC4-1.0.1 -- Are we there yet? [01:37] talon: hmm, want to check something? [01:40] sure. [01:40] _shur1 (~shushushu@vserver.electronicbox.net) left irc: Quit: changing servers [01:40] guess I found it ;) [01:41] starting to run out of steam though. [01:41] Action: talon wonders how many people actually used the bme patch before. [01:42] well, simple, they didn't ;) [01:42] for my purpose it was sufficient, because I do not ever unmount the ro bind mounts ;) [01:42] JonB (~NoSuchUse@129.142.112.33.ip.tele2adsl.dk) left irc: Quit: Leaving [01:43] yep, seems fixed now ... [01:43] seems liek a natural thing to do after your done with them. [01:43] fs/namei.c 1048,3-17 [01:43] - return error; [01:43] + goto exit; [01:43] at least when your using them to share vserver files. [01:43] _shur1 (~shushushu@vserver.electronicbox.net) joined #vserver. [01:44] first assumption about missing put() was correct ... [01:45] talon: guess I have to update the Hall of Fame entry soon ;) [01:45] at the speed you find the bugs ... [01:45] god only knows what else i will find tomorrow. [01:46] .. vserver should be bug free in two weeks ;) [01:46] i hope so. :) [01:46] and I appreciate it, and I mean it ... [01:47] so whats teh count so far? [01:47] as far as bugs. [01:47] i count at least 3. if you count getting util-vserver compiled on sparc64 [01:47] well, three different quota bugs, one of them being a general issue [01:48] the bme bug can't be accounted for vserver, as this is no vserver patch ... [01:48] iirc there where 2 userspace issues you found with enrico right? [01:48] i consider it related since thats what i use it for. [01:49] not sure if it was 2. [01:49] well only enrico knows , so the sparc64 is fix, because if, I add the HoF entry immediately ;) [01:49] yeah i think there were two bits we changed. [01:50] rob shoudl be back tomorrow or the day after btw. [01:50] before I forget ... [01:50] i think hes getting on teh train tonight. [01:50] so you should have teh sparc box up soon. [01:51] well if it's only a question of 'rob comming back' then it shouldn't be any issue to consider this done, right? [01:52] yeah shouldnt take long at all onces hes here to get you access to it. [01:52] would you like some company name associated? [01:52] hes probably goign to crash for a day though after he gets off teh train. [01:52] Bertl: Amoebasoft [01:53] okay, you got some Amoebasoft details? [01:53] and some sparc details too if possible? [01:53] nothign written down at the moment. could probably get trinity to write somthing up to date. [01:53] what details do you need for the sparc? [01:54] well, you see the entry at the top of the HoF? [01:55] its an Ultra10 with a 440Mhz UltraSPARC IIi with 256MB of ram i think an 8 gig IDE drive. running gentoo linux. [01:58] trinity is writing up a new description. [01:58] ah okay ... [01:58] Amoebasoft is a developer of advanced software and hardware [01:58] solutions for business. Utilizing current standards and technology, [01:58] they make ideas reality, and help companies achieve their [01:58] pressing business goals. [01:58] that is what I found ;) [01:59] well thats a generic description. [01:59] its a bit too much market speak for my tastes. [02:00] i was hoping maybe trinity could wirte somthign mroe down to earth. [02:00] ;) whatever you prefer ... [02:12] hmm, I guess I add update_atime() support and releas a new bme ... [02:30] hmm probably hold off until we get it online im sure to have somthign by then. [02:31] trintiy told me he wants to come up with a different company name anyway. [02:31] okay ... I'll make a not for now ... [02:31] s/not/note/ [02:31] I really have to refill that 'e's [03:00] hmm, okay that was easy, noatime and nodiratime should now work ... let's see if it compiles ;) [03:22] looks like it is working as expected ... [04:03] okay, I'll call it a day, cu all tomorrow ... [04:03] Nick change: Bertl -> Bertl_zZ [04:17] samsilvester (~sam@office1.adl.e-access.com.au) joined #vserver. [04:24] noel- (~noel@pD9FFA4B2.dip.t-dialin.net) joined #vserver. [04:31] noel (~noel@pD952CD82.dip.t-dialin.net) left irc: Ping timeout: 504 seconds [05:25] loger joined #vserver. [06:12] mmm [06:49] Frank00polo (~franko@4.13.67.211) joined #vserver. [07:06] Frank00polo (~franko@4.13.67.211) left irc: Quit: ChatZilla 0.9.52B [Mozilla rv:1.6/1] [07:32] Hey can anybody tell me if I should be worried about getting unresolved symbols when I do a depmod -a? I've just recompiled 2.4.24 with vs1.25 [07:55] riel (~riel@riel.netop.oftc.net) left irc: Ping timeout: 483 seconds [08:08] surriel (~riel@imladris.surriel.com) joined #vserver. [09:31] samsilvester (~sam@office1.adl.e-access.com.au) left irc: [10:13] Panter (~panter@p50902B51.dip.t-dialin.net) joined #vserver. [10:13] surriel (~riel@imladris.surriel.com) left irc: Ping timeout: 488 seconds [10:14] surriel (~riel@imladris.surriel.com) joined #vserver. [10:37] surriel (~riel@imladris.surriel.com) left irc: Ping timeout: 488 seconds [10:38] surriel (~riel@imladris.surriel.com) joined #vserver. [10:42] Panter (~panter@p50902B51.dip.t-dialin.net) left irc: Ping timeout: 480 seconds [10:52] Medivh (ck@62.93.217.199) got netsplit. [10:53] Medivh (ck@62.93.217.199) returned to #vserver. [10:56] stone (foobar@9593B59FDFCFE9EF220ACDF5A02394B4.lnx.nu) joined #vserver. [10:56] Nick change: Bertl_zZ -> Bertl [10:58] hi stone! [11:02] hi there.. [11:03] Action: stone is a debian developer in "love" with vserver :) [11:03] hmm, is it critical? ;) [11:08] hmm.. my wife is jealous... [11:09] sounds good 8-) debian caused some troubles in the last months, maybe you could help a little there? [11:09] I just joined to my bnc... what's the debian problem? [11:10] mainly non up to date versions, or special adapations which break with newer tools/kernels .. etc [11:11] Bertl: sarge works fine with the development and stable version for me. Woody doesn't build the development kernel/tools, only stable. [11:11] And from what I've heard, sarge is soon to become the new stable version, so I shouldn't worry about woody too much. [11:11] recently we discovered, that the debian 'version' of vserver got the networking stuff wrong, entirely my fault, I have to admit ;) [11:12] The vserver package or? [11:12] the kernel patch ... [11:12] Ahh. [11:12] morning by the way 8-) [11:12] Oh well, I build my own kernel anyhow. :) [11:12] Good morning. [11:13] Just arrived at work after being 1 hour in traffic jams. :| [11:13] most debian users switched to a non-debian kernel, but others had problems [11:14] Maybe somebody should put a pre-compiled version with the .debian configs online? [11:16] So the debian users won't have to mess with building their own kernel. [11:22] hmm, well I don't like precompiled stuff, and I guess debian users do not like it either, but I guess it would be sufficient, if the debian packages where up to date and tested ... [11:24] Well, theorectically it should be possible to build an entire debian machine without any compiler. [11:25] let me put it this way, I would not install a binary only kernel on a linux machine ... [11:25] Precompiled != binary-only. [11:25] how would you verify a precompiled kernel? [11:25] Take the source package and build it? [11:26] Or just check the gpg signature. [11:26] and then? complain about differences? [11:26] Yeah, there shouldn't be any differences if they're build on the same. [11:26] hah! simple change in compiler or binutils, completely different code ;) [11:26] So you don't install any binary packages? [11:27] well, yes I install them, after I rebuilt them ;) [11:27] I trust gpg. :) [11:27] in gpg we trust ;) [11:27] I'm unwilling and uncapable of reading any and every line of code that goes into my system. [11:28] that is probably right, but the kernel? [11:28] s/that goes into my system/of my kernel/ [11:28] :) [11:29] And I trust the stuff that comes from linux-vserver.org and vserver.13thfloor.at. (And yes, I trust dns.) [11:29] Since I don't have enough time for paranoia I'm willing to take my chances. :) [11:29] kestrel (athomas@home.swapoff.org) joined #vserver. [11:29] re [11:30] hi kestrel! [11:30] hey herbert :) [11:31] brb.. [11:34] it's time for a new devel release, don't you think so? [11:36] why the hell not [11:41] Bertl == Herbert :) Aha! [11:42] Then I should take the oppurtunity to thank the developers of Vserver for a great job. [11:43] Vserver plus LVS is two projects I really like and use alot.. [11:43] really, so you combine those? [11:43] Bertl: I have tried but don't run it now.. [11:44] Bertl: I was trying to do a "poor mans" "cluster" :) [11:44] Bertl: When I get the time I will try it again.. [11:44] I thought about that some time ago, it might have some advantages ... especially with failover vserver scenarios [11:45] yeah [11:45] keepalivd + lvs + vserver.. [11:46] It would make life mush easier with upgrades of software etc.. [11:46] something like that, I heard of some folks doing vserver failover, and thus having no downtime when updating the kernel for example ;) [11:46] just to failover to v1 and upgrade the vserver and failover to the upgraded node with v2 and so on.. [11:46] exactly this way ... [11:47] But you need 2 diffrent boxes if you want to upgrade the kernel without downtime :) [11:47] that is true [11:48] (I have read something about someone experimenting with no-reboot-kernel-upgrade) [11:48] well the kernel update per se isn't that complicated without reboot ... [11:49] the processes and the kernel memory is the tough part ... [11:50] Time for some coffe.. bbl.. [12:42] AHTOH (~Anton@212.1.230.115) joined #vserver. [12:42] hi Anton! [12:42] hi Bertl [12:42] no new devel releases? [12:43] 09:35 < Bertl> it's time for a new devel release, don't you think so? [12:43] so I'm on it ;) [12:44] ah good :) [12:45] anything special you are waiting for? [12:45] no just bugfixes, stability etc :) [12:45] got no bug reports regarding 1.3.6 yet ;) [12:45] maybe quota but that is less important for the moment [12:46] im talking about that exploits :) [12:46] ah, i see ... [12:47] i'm ready to start some servers using devel vserver+ stable utils [12:48] what about....memory limits!!! [12:48] :) [12:48] you mean enforced RSS or what? [12:52] YES! :) [12:55] hmm, you have patches which handle RSS limits on 2.4.25? or are we talking about 2.6 here? [12:56] 2.4.25 [12:56] i asked you about it the other day and you said there weren't any current patches... [12:57] Bertl: where can I find the debian patches? I can't find them through http://www.13thfloor.at/vserver/s_release/overview/. [12:59] http://vserver.13thfloor.at/Experimental/patch-2.4.24-1-vs1.24.1.diff [13:02] Ok, thanks. [13:02] np [13:13] bertl -- is lsxid problem fixed in 1.3.6.3 version? [13:14] hmm, please refresh my memory, what was the issue again? [13:14] (because it sounds very quota related ;) [13:15] gmmm i remeber we started two contextes 200 and 300 and vreated .tmp/x and /tmp/y [13:15] than 0 and 0 lsxids for them were only after reboot [13:15] yeah, that was the quota stuff, that should be fixed in q0.13, which will be out soon ... [13:15] you said me to comment me some line in kernel/inode.c and that fixed [13:16] exactly ... [13:16] what changed from 3 to 4 susubversion of 1.3.6 [13:16] ? [13:16] hmm, could not be much, probably the network stuff ... [13:17] 1.3.6 -> 1.3.7 has minimal changes ... [13:17] but that is a good sign, as we are approaching 1.4 ;) [13:18] :) [13:23] it's interesting that most people want to limit the RAM for a vserver, where the I/O system is much more critical, but nobody mentioned any limitations there yet ... [13:25] Bertl: because it's easier to sell "you have 256MB RAM" than it is to sell "you have 5MB/s disk i/o"? [13:25] hmm, good point, but in this case, it would be a feature to sell a vserver with 2GB instead of 256MB at the same price, right? [13:26] really needed a limitation on a vm -- a failed to make that work [13:26] Bertl: I don't really understand what you mean? [13:27] AHTOH: hmm, 1.3.x has vm limits, they should work as expected?! [13:27] Bertl: and for most uses and purposes disk i/o isn't the limitation, but memory is. [13:27] (And cpu). [13:28] Zoiah: was a kind of bad joke ... I meant the 'missing' limitation can be seen as 'more' memory ... [13:28] that logic means we dont need limitaions [13:28] Bertl: ahh, heheh, that confused me. : [13:30] guess we still have problems to define a good way how to 'handle' RSS limits ... or what the provider wan't to sell as 'RAM' for a vserver [13:30] s/wan't/wants/ [13:39] well, speaking for myself, the absolute pinnacle would be when you type "free" it reports a limited amount of memory and a limited amount of swap [13:39] and, of coure, those limits were enforced for processes [13:40] hmm, the fake display is easy, but how to enforce them? [13:40] yeah, that's the question alright [13:41] I'm not talking about a process consuming some memory vm/rss .. what about caches and shared stuff? [13:41] yes, exactly [13:42] if you started tracking cache usage per vserver, that would get nasty...ugh [13:42] kestrel: hmm... imho it would be nicer to have guaranteed minimum ressource available to each vserver instead of limiting the maximum ressources [13:43] and it would be kind of defeating the purpose, because if I do not allow a server to use shared caches, then you would end up with doing full partitioning like Xen ... [13:43] well, i wasn't thinking explicitly of a maximum, more of a "this is how much you have" [13:43] so it's a minimum and a maximum [13:44] herbert: that would not be good [13:44] minimum guaranteed is just bookkeeping for the admin ... [13:44] so that under high load each vserver still has some resources left, but when there's little activity on the majority, the remaining vservers grab whatever is left [13:45] s/grab/can grab/ [13:45] hmm.. any comments on the mail by Michael Hilscher (11:14am as of today) [13:45] tracking shared cache usage would be difficult, i imagine..? ie. if a context touches a cached page, it "owns" it and it is taken into account for its memory quota [13:45] i'm confused if i should update my 1.25 to 1.26 ;> [13:46] or if there will be a 1.27 any minute anyway.. [13:46] theseer: that is a bit of a concern [13:50] hmm, well, let us, once again verify that ... [13:50] TheSeer: i would've answered on the list, if i'd know the exact backgrounds, but as i see it he's missing the chattr +t call and has done a chmod +t instead [13:51] by backgrounds i mean what that chattr call does ;) [13:51] ahh, good spotting, I should actually read the mails more carefully [13:52] ah.. :) [13:52] so rebooting to 1.26 with chattr +t done should be safe for a while? [13:52] i don't like to reboot that box on a daily basis ;> [13:52] even setting the immutable flag on the /vserver dir is safe for some while ;) [13:53] hehe.. okay :) [13:54] as some guy already worte on the list, there should be a howto on what to do exactly to create a secure environment [13:54] hopeless 8-) [13:54] even though my setup seems to work nicely, i have hardly any clue if it is actually secure *g* [13:54] IIRC I posted something regarding security, and it should be available on the wiki too [13:55] TheSeer: we considered the 000 barrier secure for some years ;) [13:55] *g* [13:55] what about the /proc stuff? [13:55] is it true to say that if you are using xfs, it is impossible to secure vserver? [13:56] nope, the 1.3.7 devel release will contain xfs barrier stuff ... [13:57] Bertl: what do you do for a living anyway? I do support a lot of OS based projects.. but i can only dream of having the time to support AND develop such a big (?) thing.. [13:57] ..while still having some time left to actually make some money at a day-job [13:57] well, I do consulting from time to time ... [13:58] and I do not need much for myself ... [13:58] judging from the kernel-knowledge you must have, you prolly are a good consultant ;) [13:58] too good, nobody needed more than 5-6 hours ;) [13:58] hehe [13:59] i know that problem ;) [13:59] i should sell windows boxes [13:59] at least they require maintenance ;) [13:59] couldn't sell something I do not trust nor believe in ;) [14:00] me neither.. that's the problem.. [14:00] so i keep installing linux and hardly have to touch the box again :) [14:00] i just recently updated a RH5.2 box i basically forgot about [14:00] it just kept working [14:01] cd ..: Permission denied [14:01] chmod: Operation not permitted [14:01] Exploit seems to work. =) [14:01] [root@XXXX /]# [14:01] [root@XXXX /]# ls / [14:01] okay, verified once again vs1.26 is secure (at least agains that exploit) [14:01] bad [14:01] good [14:01] no, good ;> [14:01] hehe [14:03] Doener: do you want to answer that poor guy, or shall I? [14:03] i can do it, but i guess i can't be wrong if he is told what that chattr call is good for, and i don't know ;) [14:04] s/guess i/guess it/ [14:04] hmm, that is easy, man chattr ;) [14:05] it tells me that files won't share their last block with other files anymore, but that doesn't make much sense to me in that context [14:05] to me neither actually [14:06] you where talking about what chattr does, not what the +t is for, right? [14:06] or is the attribute just 'borrowed' for vserver specific stuff? [14:06] well, i actually meant that specific call in the context of the exploit [14:06] well, it is not borrowed, it was simply abused, thanks for that go to Sam Vilain ,... [14:07] because of compatibility reasons, we didn't change it until now ... [14:08] 'T' looks like a better fit [14:08] semantically, at least [14:09] hmm, well we have a BARRIER flag and an IUNLINK flag in 1.3.7, both not visible via lsattr/chattr [14:09] Medivh (ck@62.93.217.199) got netsplit. [14:09] Medivh (ck@62.93.217.199) returned to #vserver. [14:09] even better [14:13] ok, replied... [14:14] too late ;) [14:14] well, it won't hurt ... [14:15] guess so... [14:16] sounded to me like you would prefer me to answer ... [14:16] and to me, it sounded like you would prefer me to read the man-page and answer myself ;) [14:18] it's no problem for me, if it isn't one for you ... [14:18] Topic changed on #vserver by Bertl!~herbert@MAIL.13thfloor.at: http://linux-vserver.org/ || latest stable 1.26, devel 1.3.7, exp 0.06 [14:19] surely not, and reading man-pages can't be wrong anyways [14:24] @ALL: by the way, what is your opinion regarding FreeVPS [14:25] ? [14:27] wait a sec', i need to refresh my memory [14:27] i wonder if freevps has that vulnerability [14:28] nope, should not, as they use private namespaces ... [14:29] the feature set is nice, but it's supported platform is too limiting [14:30] although, vserver appears to be converging, feature wise [14:31] do you have any references for freevps use of private namespaces? [14:31] are you reading the mailing list? [14:31] the vserver one, yes [14:32] well, every second answer from Alexey, is ' a better solution is private namespaces, as FreeVPS does it' or 'that would not happen with private namespaces' ... [14:34] hmmm [14:34] the other answers are, 'we have virtualized network, that is the only solution' ;) [14:34] heh, yeah, i've read that [14:36] he's persistent [14:38] i wonder what he means [14:40] arekm (misiek@ikar.t17.ds.pwr.wroc.pl) left #vserver. [14:42] herbert, did you decide whether to do a "simple" port of vserver to 2.6, or are you going to do a full port to all the new 2.6 featuers? [14:43] yes, the decision was simple, after the replies I got on the ml ... [14:44] cool [14:44] there will be a 'final' 1.4 for 2.4.x, and two branches for 2.6 one, which does it the way it is done now, and the other which will favor SE and other 'new' 2.6 stuff ... [14:45] very cool [14:45] we'll see, probably one has a better takeup than the other ... [14:45] but I assume both branches will coexist for a while ... well, actually the SE branch will need some time to start off first ;) [14:46] heh, yeah [14:47] has anybody downloaded the 1.3.7 patch yet? [14:48] i have not [14:48] i can try it at work on my sparc though [14:48] the two other vservers i run are production, so... [14:48] just locked down the download, because I discovered that the patch was incorrect ... [14:49] or to be precise, incomplete ... [14:51] i wonder if freevps is vulnerable to that recent kernel vulnerability...it's based on 2.4.18 [14:51] probably ... [14:52] I think the proc security is also an issue there ... [14:52] don't know if they fixed the stability issues ... [14:52] mmm, yes [14:54] i wonder if anybody has done vserver migration with shared disk of some description [14:54] scsi, or fc [14:54] that would be cool [14:56] kestrel: I can't see why it wouldn't work. I have a bunch of MSA1000 stuff at work but that's all used in production. :| [14:57] damn production, can't be used for anything, keeps everybody fromt esting 8-) [14:57] okay, 1.3.7 is corrected ... [14:58] Even though some of the stuff isn't even is used, but has to be kept in stand-by permanently. :) [14:59] yeah, it would be cool to play with it [15:01] surriel (~riel@imladris.surriel.com) left irc: Ping timeout: 488 seconds [15:15] okay, everything should be fine now, quota will follow this evening, I guess, now it's time for siesta ;) [15:16] have a good snooze herbert [15:19] hmm, not yet, another 'exploit still working' [15:21] they're coming thick and fast [15:55] click (click@gonnamakeyou.com) joined #vserver. [15:56] hi click! [15:56] damn, I never remember to add this chan and network to autorejoin [15:56]