1296432770 Q * manana Ping timeout: 480 seconds
1296433401 M * Guy- cehteh: I have old xfs filesystems too, and they aren't slow either
1296433423 M * Guy- I very occasionally run the xfs defragmenter (about once every two months
1296433425 M * Guy- )
1296433457 M * Guy- cehteh: xfs is slow if you put it on raid5 or raid6 without delaylog and without an external journal
1296433474 M * Guy- (even then, only metadata updates are slow)
1296433823 J * manana ~mayday090@84.17.25.149
1296434439 M * Bertl_oO off to bed now ... have a good one everyone!
1296434444 N * Bertl_oO Bertl_zZ
1296434574 Q * Janno Quit: ZNC - http://znc.sourceforge.net
1296434923 J * Janno Janno@server3.raumopol.de
1296435535 Q * rgl Ping timeout: 480 seconds
1296437277 Q * Piet Remote host closed the connection
1296437322 J * Piet ~Piet__@04ZAACD02.tor-irc.dnsbl.oftc.net
1296441329 Q * Huendchen Ping timeout: 480 seconds
1296441345 J * Huendchen ~YOSEF@ip-95-223-36-117.unitymediagroup.de
1296444785 Q * swenTjuln Ping timeout: 480 seconds
1296445622 Q * hparker Quit: Quit
1296447769 Q * Piet Remote host closed the connection
1296447814 J * Piet ~Piet__@04ZAACD2B.tor-irc.dnsbl.oftc.net
1296456686 J * ghislain ~AQUEOS@adsl2.aqueos.com
1296457760 Q * FireEgl Read error: Connection reset by peer
1296458793 J * swenTjuln ~kvirc@77.111.2.36
1296459220 Q * Piet Ping timeout: 480 seconds
1296459854 J * Piet ~Piet__@04ZAACD41.tor-irc.dnsbl.oftc.net
1296461163 J * FireEgl ~FireEgl@173-25-19-139.client.mchsi.com
1296462493 N * Bertl_zZ Bertl
1296462497 M * Bertl morning folks!
1296462521 M * arekm morning
1296464313 J * petzsch ~markus@dslb-088-075-167-242.pools.arcor-ip.net
1296465127 J * kir ~kir@swsoft-msk-nat.sw.ru
1296465369 P * kir 
1296470742 M * Bertl arekm: http://vserver.13thfloor.at/ExperimentalT/delta-cred-feat05.diff
1296471148 M * arekm testing
1296471289 M * Bertl tx
1296474948 M * Huendchen what is backup for restore my hostsystem?
1296475009 M * Bertl whatever works for a normal Linux system will work for a Linux-VServer host
1296475034 M * Bertl e.g. dump/restore for ext2/3, rsync, tar (proper options)
1296475203 M * Huendchen ?! i need only backup the vservers i think /etc/vservers and /var/lib/vservers ?
1296475371 M * Bertl if you just want to have a backup of the guest systems, then yes, but you asked for the host system
1296475482 Q * petzsch Quit: Leaving.
1296476551 J * petzsch ~markus@dslb-088-075-167-242.pools.arcor-ip.net
1296478663 M * nkukard Bertl, what time is daniel_hozac normally online?  I"m still looking for the latest yum patch that isn't in the latest tarball
1296478680 M * daniel_hozac always check the svn repo.
1296478692 M * daniel_hozac ensc added a patch a few days ago.
1296478707 M * nkukard thanks daniel_hozac , busy checking now
1296478840 M * nkukard daniel_hozac, latest yum is 3.2.29 and the 3.2.27el6 doesn't apply to it :(
1296479296 M * nkukard daniel_hozac, I tried to modify it a bit to get it working ,but to no avail, my python skills are not that spectacular
1296481154 Q * arekm Quit: reboot
1296481902 J * arekm matrix157@83.238.65.58
1296481911 M * arekm Bertl: heh, exploded in my case, http://pld.pastebin.com/yQ8i4H6S
1296481922 M * arekm Bertl: don't know why yet
1296482441 M * daniel_hozac kmemcheck?
1296482461 M * daniel_hozac doesn't look like something you want enabled in a production kernel.
1296483368 M * arekm it's not production, it's test kernel for testing credentials
1296484002 M * Bertl looks like that happened long before the code in question got touched
1296484020 M * Bertl (but I can't tell without a complete bootup log)
1296484088 M * arekm ipmi sol isn't perfect; trashes boot logs often
1296484102 M * Bertl get a serial console :)
1296484117 M * Bertl anyway, looks like a bad acpi bios to me
1296484165 M * arekm machine is 300km away, so a bit hard to get anything right now hehe
1296484257 M * Bertl tryp booting with acpi=off
1296484311 M * Bertl if that doesn't work, try with noacpi instead
1296484430 N * ensc Guest2246
1296484440 J * ensc ~irc-ensc@p5DF2E797.dip.t-dialin.net
1296484503 M * arekm it booted before that little delta, weird 8-) trying various things anyway
1296484599 Q * Guest2246 Ping timeout: 480 seconds
1296484657 M * Bertl could be a race condition, could be the patch .. but as I said, it doesn't look like the patched code was involved ...
1296484664 M * Bertl nap attack ... bbl
1296484669 N * Bertl Bertl_zZ
1296485955 M * arekm Bertl_zZ: yee, works!
1296486150 M * arekm Bertl_zZ: tested with python script in a few ways + with forking few thousands processes after vserver enter; exit; enter etc
1296486273 J * arekm_ arekm@carme.pld-linux.org
1296486273 Q * arekm Quit: leaving
1296486277 Q * arekm_ 
1296486280 J * arekm arekm@carme.pld-linux.org
1296486901 N * ensc Guest2252
1296486911 J * ensc ~irc-ensc@p5DF2EF31.dip.t-dialin.net
1296487156 Q * nkukard Ping timeout: 480 seconds
1296487311 Q * Guest2252 Ping timeout: 480 seconds
1296487331 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647
1296488349 J * bsingh ~balbir@122.172.7.42
1296488580 J * yarihm ~yarihm@access.zrhcz.openbox-ict.com
1296488617 J * dowdle ~dowdle@scott.coe.montana.edu
1296490046 J * lp ~me@193.159.185.162
1296490057 M * lp hey there
1296490059 M * lp :)
1296490231 M * lp is it possible to run a vserver inside a vserver, or at least a lxc in a vserver?
1296490240 M * daniel_hozac no.
1296490269 M * lp sad story :)
1296490306 M * lp OK but thanks, this question seemed to be answered to 100%
1296490384 M * lp have a nice day & thanks for the great project
1296490386 P * lp 
1296491998 Q * yarihm Ping timeout: 480 seconds
1296492060 J * yarihm ~yarihm@access.zrhcz.openbox-ict.com
1296492167 Q * yarihm 
1296492290 Q * bsingh Read error: Operation timed out
1296492392 Q * ntrs Remote host closed the connection
1296492475 J * ntrs ~ntrs@209.135.140.188
1296493072 J * bsingh ~balbir@122.172.2.80
1296493563 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1296495406 Q * petzsch Quit: Leaving.
1296497202 J * dna ~dna@dslb-088-074-208-135.pools.arcor-ip.net
1296497328 Q * bsingh Read error: Operation timed out
1296497994 J * bsingh ~balbir@122.172.17.28
1296498521 N * Bertl_zZ Bertl
1296498552 M * Bertl arekm: so for now the best solution, yes?
1296498571 M * Bertl arekm: what we need to check/verify is that the cred accounting isn
1296498585 M * Bertl 't messed up, and we do not leak credential somehow
1296498588 M * Bertl +s
1296498656 M * Bertl i.e. try to enable the credential debugging stuff, if not done so already, and try to get a few 10000 start/stop cycles for a test guest (best make it start and stop quite fast)
1296498684 M * Bertl also test with and without hitting that guest with enters
1296498718 M * Bertl monitor kernel memeory and the credential debug stuff during this test
1296498761 M * Bertl I'm optimistic that we got the proper solution, it took me some time to figure a way which isn't too intrusive and doesn't copy existing kernel code
1296499458 M * arekm I assume credential debug is just printk? (enabled but haven't seen any messages in dmesg)
1296499656 M * arekm Bertl: for now I didn't hit any issue like fork() - testing the same way on previous (.36) patches triggered fork() problems quite fast
1296499668 M * arekm and SO_PEERCRED returns proper values in all fields
1296500033 M * Bertl okay, I'll add it to the next 2.6.38-rc and for 2.6.37 as well
1296500107 M * arekm I'll put prebuilt kernel for pld users, so they will also test it (when of course .37 patch will be available)
1296500572 J * ktwilight ~keliew@91.176.247.205
1296500607 M * Bertl shouldn't take too long to adapt, AAMOF, it might already apply
1296501150 M * Bertl daniel_hozac: what's your opinion on the patch?
1296501241 Q * bsingh Ping timeout: 480 seconds
1296501249 M * Bertl arekm: (yep, patch -l -p1 should be fine for 2.6.37 as well)
1296501570 Q * Piet Ping timeout: 480 seconds
1296501819 J * bsingh ~balbir@122.172.14.145
1296502257 J * Piet ~Piet__@04ZAACEQ0.tor-irc.dnsbl.oftc.net
1296502367 M * arekm you mean .38 + feat apply to .37?
1296502383 M * Bertl I mean, feat05 applies to latest .37-rc
1296502424 M * Bertl but I'll upload a new .37 patch in a few minutes
1296502643 M * arekm cool
1296503049 M * Bertl new 2.6.37 patch is up, have fun with that gigantic patch :)
1296503276 M * arekm great, thanks!
1296503287 J * petzsch ~markus@dslb-088-075-167-242.pools.arcor-ip.net
1296503625 Q * hparker Quit: Quit
1296504197 J * slamczyk ~slamczyk@CF3840C2.endinfosys.com
1296504272 M * slamczyk Anyone know if there is a way to hard allocate memory using cgroups?
1296504376 M * slamczyk I know I can hard allocate processor, but I want to know if there is a way to hard allocate memory using cgroups.
1296504440 P * slamczyk 
1296504447 J * slamczyk ~slamczyk@CF3840C2.endinfosys.com
1296504742 P * petzsch 
1296504784 M * Bertl slamczyk: what do you mean by 'hard allocate'?
1296504820 M * slamczyk I need to make sure one machine gets a dedicated 24GB or ram, while the remaining 8GB I want to share across four other machines.
1296504852 M * Bertl cpusets might help you there
1296504887 M * slamczyk That make sense?  So like with CPU's I can tell it to use 0-7 for the first machine and the rest of them just put 8-15 in and all four machines will have the second set of processors while the first machine gets the dedicated first 8.  I need to do something similar with memory
1296504986 M * Bertl cpusets can control memory as well
1296505335 M * slamczyk Looking that over, I think I can share memory, but not actually hard allocate it.  Saying like guest 1 = first 24GB, then guest 2-5 = last 8GB.
1296505398 M * Bertl we are back to what 'hard allocate' means for you
1296505448 Q * cuba33ci Read error: Connection reset by peer
1296505474 M * slamczyk In my case it is simple.  I need the first guest to have 24GB of unshared memory, while the other four guests can share the last 8GB of ram.
1296505512 M * Bertl so? make the memory set for the first guest exclusive and don't use up resources on the host
1296505567 M * slamczyk The first part of that is correct, I want the first guest to have 24GB of exclusive memory.  The other four guests I want to share the remaining 8GB or RAM.
1296505568 J * cuba33ci ~cuba33ci@111-240-167-72.dynamic.hinet.net
1296505643 M * Bertl putting the other guests in a separate set should work, no?
1296505677 M * Bertl (assuming that your setup/system can allocate the memory sets at all)
1296505684 M * slamczyk I mean, that sounds right, but maybe I am missing what you are saying.  I was looking into cpuset, but all I can find is the cgroup page, am I right?
1296505716 M * Bertl http://lwn.net/Articles/127936/
1296505719 J * rgl ~Rui@a89-154-33-145.cpe.netcabo.pt
1296505724 M * Bertl cpusets are a lot older than cgroups
1296505767 M * slamczyk Oh yeah we are on the latest kernel I believe, my co-worker is in here alot asking questions, and he set it all up.  He knows way more than me about this stuff.
1296505817 M * slamczyk  2.6.36-vs2.3.0.36.36
1296505822 M * slamczyk Is the kernel we are running
1296505831 M * Bertl good
1296505878 M * slamczyk Way beyond me is what I am thinking
1296505880 M * slamczyk :)
1296506290 J * ryker ~ryker@199.117.46.145
1296506666 M * ryker hi Bertl.  I work with slamczyk and just trying to understand what he just told me about memory allocation.  We are using cgroups on a custom compiled 2.6.35 vserver kernel.  We can't use old style cpusets, can we?
1296506712 M * ryker the primary goal here is to allocate 24G of RAM to one guest and not have the other guests ever be allowed to touch that 24G of RAM.
1296506747 M * ryker So, even if the other guests end up causing the server to use swap space, we want that 24G of real RAM still available to the one guest.
1296506754 M * ryker Does that sound feasible?
1296506756 M * Bertl that's probably hard to accomplish with cgroups atm
1296506854 M * ryker but it was possible with old style cpusets I think, right?
1296506878 M * Bertl as I said, if cpusets work on your setup (btw, 2.6.35 is not the best choice) then you might be able to do that
1296506881 M * ryker by old style, i mean what was available with the vserver kernel before we had cgroup support
1296506898 M * ryker what's wrong with 2.6.35?
1296506922 M * ryker it was the latest kernel available a couple months ago.  i guess it's time to compile a new one
1296507026 M * Bertl IMHO 2.6.36 is the first useable kernel since 2.6.22
1296507031 M * ryker wow, i didn't realize 2.6.38 is almost released already
1296507054 M * Bertl yep
1296507063 M * ryker we really haven't had any issues with it.  luckily I guess from what you are saying.
1296507065 Q * Piet Ping timeout: 480 seconds
1296507154 M * Bertl 2.6.35 isn't that bad as the 2.6.26 kernel debian picked was :)
1296507197 M * ryker is there a yeah, we stuck with 2.6.18 actually until we went with 2.6.35 to get cgroup functionality
1296507227 M * ryker Bertl: you're in here all of the time.  are you a vserver dev or just an avid user/fan like I am? :)
1296507361 J * nkukard ~nkukard@41-133-198-249.dsl.mweb.co.za
1296507506 M * Bertl ryker: kind of both I guess :)
1296507856 M * ryker :)  yeah, I didn't see you on the dev page, but I wasn't sure how up to date it was
1296508004 M * Bertl it's pretty up-to-date and I'm there ...
1296508122 J * Piet ~Piet__@04ZAACEUB.tor-irc.dnsbl.oftc.net
1296508177 P * slamczyk 
1296508227 M * ryker oh, sry. i don't see your irc name on there.
1296508246 M * ryker lol
1296508249 M * ryker ok, right at the top
1296508252 M * ryker nvm, i'm blinkd
1296508255 M * ryker *blind
1296508335 M * Bertl np :)
1296508633 M * ryker Bertl: so, since I have you here, I'm wondering if we can donate some money to get some bugs fixed soon.  I talked to my manager and he is happy to do that.  Is this the appropriate place to discuss this?
1296508846 M * Bertl sure, but you do not need to 'pay' to get bugs fixed, i.e. we fix them as soon as they are reported
1296508863 M * Bertl it's different with feature requests and nice to have stuff
1296508866 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647
1296508922 M * Bertl what 'bugs' do you experience?
1296508936 M * ryker reboot doesn't work in a guest when cgroups are enabled.
1296508974 M * ryker i mentioned it on here before.  Earlier this month I think.
1296508984 M * ryker is there a bug tracker I can submit this to?
1296509009 M * Bertl yes, we have one, but here is better/faster, can you provide a guest and setup which shows this misbehaviour?
1296509015 M * ryker i tracked it down to here: /usr/lib64/util-vserver/vserver.functions +1525
1296509064 M * ryker sure.  I assume you need root access to the host and guest?
1296509091 M * Bertl no, not really, actually I'd prefer a guest to download (tar, dump, etc) and the according config you use
1296509118 M * Bertl (but if that is not practical, I can check on your host as well)
1296509162 M * ryker Bertl: oh, ok.  np.  Do you want me to upload it somewhere or provide you a link to download?
1296509185 M * Bertl whatever works for you, you can msg, me in private if that helps
1296509246 M * ryker no bugtracker though.  seems odd.  do I need to donate a server for you guys to use a bugtracker/ticket system?  ;)
1296509341 M * Bertl as I said, we have one, but it isn't really used/necessary
1296509523 M * ryker yeah, i was just giving you a hard time about that. :)  Thanks for the offer to help with this issue.  I'll get together a tarball and pm you a link to download it.  I'm still going to tell my manager we need to donate anyway. :)
1296509600 M * ryker hah, yeah, we are using 2.6.36, not 2.6.35.  I remember I did have some issues with 2.6.35, but it may have just been a problem with the features I compiled in.
1296509646 M * Bertl donations of all kinds are always welcome ...
1296510078 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com
1296510362 Q * BenG 
1296510765 Q * bonbons Quit: Leaving
1296512001 Q * dna Quit: Verlassend
1296513841 Q * bsingh Ping timeout: 480 seconds
1296514395 J * bsingh ~balbir@122.172.17.162
1296515731 Q * bsingh Ping timeout: 480 seconds
1296516268 J * bsingh ~balbir@122.172.30.31