1306454524 Q * Piet Ping timeout: 480 seconds
1306454680 M * Bertl hi
1306454744 N * Piet_ Piet
1306456108 M * er hi Bertl. hope all's good with you.
1306456343 M * Bertl everything fine, thanks, and you?
1306456400 M * er not bad
1306456405 M * er I mean all good
1306456740 M * er are you aware of any issues on 2.6.32 kernels that might affect copy-on-write in guests?
1306456850 M * er we have recently upgraded to 2.6.32.71 with vserver 2.3.0.36.29.6. since we did that, writing to hardlink-populated files in guests no longer unlinks them, but writes into the original file in the source reference image.
1306456869 M * er we didn't have this issue on our previous 2.6.27 kernel
1306456930 M * er CONFIG_VSERVER_COWBL=y 
1306456967 M * er [root@planetlab-01 etc]# showattr /vservers/yale_p4p/etc/sudoers 
1306456967 M * er -----UICX /vservers/yale_p4p/etc/sudoers
1306457259 Q * dowdle Remote host closed the connection
1306457839 M * Bertl well, hard linked files without the proper flags will not be affected by CoW link breaking
1306457858 M * Bertl have you checked that the flags are intact after the kernel change?
1306457871 M * Bertl (2.6.26/27 was known to use wrong flags)
1306457918 M * er i see. the line i pasted is an example of the xflags of a file that should be Cow link broken but isn't
1306457992 M * er it has 'I' set which IIUC implies the link breaking option
1306458033 M * Bertl I is immutable, U is the Unlink Inversion
1306458103 M * Bertl no idea what the X actually is
1306458141 M * Bertl but I would suggest to move to a more recent kernel anyways, preferably a 2.6.38.x one
1306458211 M * Bertl (especially as it seems you are updating, so why not update to a good kernel :)
1306458385 M * Bertl sorry, I'm rather tired atm, so I guess I'm off to bed now ... 
1306458396 M * Bertl have a good one everyone!
1306458403 N * Bertl Bertl_zZ
1306458429 M * er thanks for the tips, Bertl_zZ. 
1306458441 M * er we will probably upgrade further, but looking for a stopgap fix.
1306461992 J * derjohn_foo ~aj@p4FFD1AAD.dip.t-dialin.net
1306462385 Q * derjohn_mob Ping timeout: 480 seconds
1306462705 Q * fback Ping timeout: 480 seconds
1306462908 J * fback fback@red.fback.net
1306463271 J * ryker ~Adium@c-76-16-115-27.hsd1.in.comcast.net
1306464175 Q * hparker Quit: Quit
1306464189 Q * er Quit: er
1306466209 J * er ~sapan@c-68-36-169-19.hsd1.nj.comcast.net
1306466249 Q * er 
1306471797 J * bsingh ~balbir@122.172.32.145
1306472235 Q * quasisane_ Quit: leaving
1306472710 Q * quasisane Quit: leaving
1306475907 Q * ryker Quit: Leaving.
1306476309 J * quasisane ~sanep@c-76-24-80-97.hsd1.nh.comcast.net
1306476992 J * ghislain ~AQUEOS@adsl2.aqueos.com
1306477730 Q * SwenTjuln Ping timeout: 480 seconds
1306477756 J * SwenTjuln ~SwenTjuln@77-111-2-36.ipv4.tusmobil.si
1306478008 Q * bsingh Read error: Connection reset by peer
1306478776 J * bsingh ~balbir@122.167.170.30
1306479573 J * ryker ~Adium@c-76-16-115-27.hsd1.in.comcast.net
1306480175 Q * derjohn_foo Ping timeout: 480 seconds
1306480592 J * yosefisoef 4751513c@ircip1.mibbit.com
1306480738 M * yosefisoef can anyone give me some help?
1306481161 Q * yosefisoef Quit: http://www.mibbit.com ajax IRC Client
1306482936 J * petzsch ~markus@dslb-092-078-118-221.pools.arcor-ip.net
1306484551 M * ghislain hopefully
1306484561 M * ghislain depend on the issue
1306484684 J * derjohn_foo ~aj@ip-81-210-228-18.unitymediagroup.de
1306486250 J * thierryp ~thierry@zanzibar.inria.fr
1306486664 J * cyril_ ~cyril@pas38-1-82-67-71-177.fbx.proxad.net
1306487264 M * cyril_ hi!
1306487463 M * cyril_ Anyone know if the debina kernel 2.6.38.3-vs2.3.0.37-rc14-beng on repo.psand.net repository can be use on production environment ? (sorry for my very bad english :/)
1306487526 M * pmjdebruijn _personally_ for production environment I tend to stick to the -longterm kernels
1306487537 M * pmjdebruijn we still on 2.6.27.x (ancient) and probably moving to 2.6.32.x (soon)
1306487746 M * cyril_ yes generally me too, but in this case the server has a network card which work only with 2.6.36+ kernel :(
1306487773 M * cyril_ so the official debian kernel (2.6.32) doesn't work
1306488269 M * cyril_ hmm there is a linux-image-2.6.36.4-vs2.3.0.36.39-beng, this one isn't on rc
1306488718 M * pmjdebruijn cyril_: build your own
1306488797 M * ghislain beng kernel  are used in production by the maker of them so generaly they are quite good anyway i do not use them myself so i cannot vouch for him
1306488812 M * ghislain building your own is not very difficult in debian
1306489498 M * cyril_ i know but i need to build a 2.6.36 kernel with vserver patch and the beng kernel is exactly what i want :)
1306490636 Q * petzsch Quit: Leaving.
1306493083 M * daniel_hozac 2.6.38 is longterm maintained as well.
1306493093 M * pmjdebruijn oh?
1306493096 M * pmjdebruijn cool
1306493111 M * pmjdebruijn but I mostly meant with a highish patch level
1306493193 M * daniel_hozac oh, so they need to have been neglected for some amount of time before they are "stable"?
1306493276 M * pmjdebruijn daniel_hozac: highest patch level is not being neglected :p
1306493369 M * daniel_hozac longterm kernels get nowhere near the same attention as the stable kernel.
1306494299 M * sannes1 If I'm going to test something after adding a vcmd, .. I saw there was a vcmd util http://linux-vserver.org/VCMD_HowTo , but not where to get it .. any ideas?
1306494811 M * fback daniel_hozac: what is the minimal version of utils for the latest vserver kernel patches?
1306494821 M * daniel_hozac latest is always best.
1306494834 M * daniel_hozac sannes1: http://vserver.13thfloor.at/Experimental/TOOLS/
1306494859 M * fback daniel_hozac: yup, but this doesn't answer the question :-)
1306494901 M * daniel_hozac sure, i simply don't keep track
1306494906 M * daniel_hozac well
1306494907 M * daniel_hozac actually
1306494911 M * daniel_hozac for the latest patch
1306494916 M * daniel_hozac you should always run the latest utils.
1306494999 M * fback latest as in "released" (from people.linux-vserver.org/~dhozac) or from the git?
1306495020 M * sannes1 daniel_hozac: Thank you again :)
1306495066 M * daniel_hozac either, typically they are the same.
1306495075 M * daniel_hozac of course, latest pre-release is a lot easier...
1306495793 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com
1306496623 J * kir ~kir@swsoft-msk-nat.sw.ru
1306496654 Q * BenG Quit: I Leave
1306496877 P * kir 
1306499689 Q * fback Remote host closed the connection
1306499893 J * fback fback@red.fback.net
1306499917 N * Bertl_zZ Bertl
1306499924 M * Bertl morning folks!
1306500637 M * sannes1 Morning Bertl :P
1306500715 M * sannes1 I have a question for you, in the vserver patch the unshare mask is called umask, but this is used by other places, would it not be a problem if I renamed it unshare_mask , or atleast usmask or something like that?
1306500742 M * sannes1 (also got the vcmds to work and do what I need)
1306501146 M * Bertl why would you want to rename it in the kernel?
1306501237 M * sannes1 Just to make it more readable ..
1306501292 M * Bertl where are the places which 'confuse' you because umask is already used?
1306501340 M * sannes1 when people refer to umask, most people refer to the file umask ..
1306501377 M * sannes1 Just test it out on some random people, I could be mistaken of course
1306501432 M * Bertl when people refer to operating system, most of them are talking about windows, so?
1306501459 M * daniel_hozac i think it should be nsmask
1306501486 M * sannes1 Yes, you are right of course, I was just thinking it would be nice if you could read and then go "ah, that is what it does".. 
1306501491 M * Bertl currently it is solely used to check if you _can_ _unshare_ a namespace
1306501527 M * Bertl so, we could call it cunsmask but I doubt that would be more descriptive :)
1306501571 M * sannes1 clone_mask since it is checked against CLONE_* .. :P
1306501643 M * sannes1 nsmask is nice too, since it tells me it has something to with namespaces ..
1306501784 M * Bertl well, I guess you can run the entire patch through sed 's/umask/nsmask/g' for your purpose
1306501793 M * sannes1 not that it matters for me right now
1306501826 M * Bertl for me, nsmask would be a lot more confusing :)
1306501957 M * Bertl anyway, if there is a reasonable large demand to change the name, I'm fine with that
1306501986 M * sannes1 Speaking of patch you want the small diff of what i've done on top of the latest or .. ?
1306502051 M * Bertl you did some kernel changes?
1306502095 M * Bertl if so, sure I'm interested to look at them
1306502482 M * sannes1 All the functions were already there, as you can see I did almost nothing: http://www.sannes.org/misc/expose_vserver_umask_set_and_get.patch
1306503279 M * sannes1 Think I almost did more to vcmd to test it out
1306503454 M * Bertl okay, will incorporate that/similar on the next release
1306503504 M * Bertl the interesting part is, I'm pretty sure we added that because somebody asked for it ... but obviously, he/she could not have actually used it :)
1306503536 J * petzsch ~markus@dslb-092-078-118-221.pools.arcor-ip.net
1306503606 M * sannes1 So now I'm going to see what I can do to util-vserver, but I actually have problems compiling it manually from the git tree :P For some reason the ./configure I end up with does not work (complaining about missing install-sh).
1306503816 M * sannes1 Bertl: Well, I asked for it quite a while back, but then it had worked (only needed namespace in ccaps), but stopped when vx_umask/can_unshare was introduced, if my memory serves me correctly... but did not have the time to look very closely at it. :P
1306504415 Q * petzsch Quit: Leaving.
1306504819 J * crystalblue66 crystal@82.78.140.99
1306504837 M * crystalblue66 www.xxxlivecam.info
1306504907 M * sannes1 daniel_hozac: What is all this V0, V1 V2 stuff? Is it backwards compability? Do I need to that for new calls?
1306504923 M * sannes1 new vcmds I mean
1306504934 M * Bertl off for now ... bbl
1306504939 N * Bertl Bertl_oO
1306504954 M * daniel_hozac when you only have one version, no, you don't need more.
1306505091 Q * bsingh Ping timeout: 480 seconds
1306505131 M * sannes1 How about the VC_CALL(CALL_VC_V11(..), CALL_VC_V13(..)) ?
1306505146 Q * crystalblue66 autokilled: This host violated network policy. Mail support@oftc.net if you think this in error. (2011-05-27 14:05:46)
1306505689 M * daniel_hozac you need whichever one is appropriate.
1306506227 M * sannes1 Okay, so in my case that would be .. syscall(__NR_vserver, 14, xid, data); ..
1306506267 M * daniel_hozac no.
1306506293 M * daniel_hozac never use syscall, never use a raw number.
1306506335 J * SwenTjuln_ ~SwenTjuln@193.164.140.194
1306506577 J * dowdle ~dowdle@scott.coe.montana.edu
1306506609 Q * SwenTjuln Ping timeout: 480 seconds
1306506609 N * SwenTjuln_ SwenTjuln
1306506851 M * sannes1 daniel_hozac: vserver(VCMD_....)  is fine to use?
1306506885 M * daniel_hozac no.yes.
1306506888 M * daniel_hozac -no.
1306506971 M * sannes1 When is it appropriate to use CALL_VC_XYZ?
1306507006 M * daniel_hozac always.
1306507012 M * daniel_hozac you might want to look at other syscalls.
1306507033 M * sannes1 Yes, I looked at ccaps f.ex. because it does pretty much what I want ..
1306507081 M * sannes1 but, I don't quite grasp what CALL_VC_V13, V13A, V23 and such do, or if I'm supposed to use them at all
1306507101 M * daniel_hozac yes
1306507106 M * daniel_hozac they always get used.
1306507122 M * daniel_hozac pick the one that is right.
1306507156 M * sannes1 So, should I just go for the one with the highest number?
1306507159 M * sannes1 since it is new?
1306507172 M * daniel_hozac whichever number is where the call was introduced.
1306507220 Q * thierryp Remote host closed the connection
1306507239 M * sannes1 So V23 is 2.3 ?
1306507284 M * sannes1 vs2.3.0.37rcX that means V23 .. that makes sense in a way :P
1306507525 M * derjohn_foo Hi, when I give ccap sys_admin to a guest, what else to I need to make losetup work in the guest? the kernel module loop is already loaded.
1306507635 M * daniel_hozac derjohn_foo: you mean bcap?
1306507674 M * daniel_hozac sannes1: V23 is just 0x00020300
1306507697 M * daniel_hozac V23P is closer, but not really enough...
1306507709 M * daniel_hozac although, for a new call, V23 is probably fine.
1306507847 M * sannes1 :) Great, and thanks again, I'll send you some patches when I get something working. I probably have some more questions :)
1306507860 M * derjohn_foo daniel_hozac, .... my collegues did that. They said ccap, but bacp seems to be more correct after rethinking ist...
1306508129 N * click_ click
1306508725 Q * derjohn_foo Ping timeout: 480 seconds
1306510261 J * bsingh ~balbir@122.172.2.125
1306511254 Q * _nono_ Quit: Leaving
1306516810 Q * ghislain Quit: Leaving.
1306524108 J * kbad ~Adium@ip-66-33-206-8.dreamhost.com
1306524156 M * kbad Is "vsched: vc_set_sched(): Function not implemented" indicative of util-vserver being out of date?  (2.6.38.7-vs2.3.0.37-rc15+, util 0.30.216-52)
1306524604 M * Marillion i think, yes
1306524632 M * Marillion vserver-info say which version has you
1306524718 M * daniel_hozac "0.30.216-52" doesn't exist.
1306524730 M * Marillion kbad: util-vserver: 0.30.216-pre2925 works fine
1306524730 M * daniel_hozac so you might want to double check on that.
1306524793 M * Marillion awesome job for developers
1306524848 M * Marillion /s/for/from
1306524916 J * bonbons ~bonbons@2001:960:7ab:0:44fa:1980:5358:6b32
1306524924 M * kbad k, lemme build that up
1306524935 M * kbad the −52 was my package iteration 
1306524936 M * kbad sorry
1306524937 M * kbad :X
1306524955 M * kbad pre2914, updating to 2967
1306525999 Q * bsingh Ping timeout: 480 seconds
1306526542 J * petzsch ~markus@dslb-092-078-118-221.pools.arcor-ip.net
1306529478 Q * kbad Quit: Leaving.
1306530309 J * kbad ~Adium@ip-66-33-206-8.dreamhost.com
1306533078 M * fback daniel_hozac: which cgroup opts are required and which are recommended / suggested ?
1306533101 M * daniel_hozac hmm?
1306533185 M * fback some parts (eg scheduler?) depend on cgroups, right?
1306533285 M * daniel_hozac vserver-stat depends on the memory cgroup-
1306533298 M * daniel_hozac everything else is just a "that's how you configure it now"
1306533299 M * fback the patch depends on namespaces, so required ones are automatically selected
1306533362 M * fback but cgroup support is not even enabled, ad iirc there are some options needed 
1306533373 M * fback thank you, daniel
1306533375 M * daniel_hozac none are needed.
1306533447 M * fback but you just said util-vserver depends on memory cgroup?
1306533471 M * daniel_hozac it's not a _strict_ requirement.
1306533475 M * daniel_hozac it works better.
1306533481 M * daniel_hozac and is default enabled if that is the case.
1306533488 M * daniel_hozac but it is not absolutely necessary.
1306533519 M * daniel_hozac vserver-stat will just revert to the old sum-of-processes.
1306533591 M * fback daniel_hozac: any other cgroups that could improve something?
1306533616 M * daniel_hozac they all give you something.
1306533627 M * daniel_hozac read the description to see if you think it might be useful to you.
1306533836 M * daniel_hozac personally i enable all of them, but don't use ns.
1306533890 M * fback daniel_hozac: by don't use you mean don't mount it?
1306533973 M * daniel_hozac right. util-vserver excludes it, since it doesn't work with it.
1306534033 M * fback ok, thank you
1306534089 M * fback I'm nearly out of batteries, see you
1306534135 Q * petzsch Quit: Leaving.
1306534245 Q * wurtel_ Ping timeout: 480 seconds
1306535673 Q * bonbons Quit: Leaving
1306536290 Q * ryker Quit: Leaving.
1306537484 Q * kbad Quit: Leaving.
1306538899 N * ensc Guest2500
1306538909 J * ensc ~irc-ensc@p5DF2E93C.dip.t-dialin.net
1306539320 Q * Guest2500 Ping timeout: 480 seconds
1306540392 Q * dowdle Remote host closed the connection
1306540544 J * Piet_ ~Piet__@04ZAABYB0.tor-irc.dnsbl.oftc.net
1306540644 N * Piet Guest2504
1306540644 N * Piet_ Piet