1399005527 M * Bertl off to bed now ... have a good one everyone!
1399005536 N * Bertl Bertl_zZ
1399010321 Q * ser Quit: Quit.
1399010325 J * ser ~ser@2610:28:3090:3001:a:a:a:beef
1399012079 Q * opuk Ping timeout: 480 seconds
1399013477 J * Ghislain ~aqueos@adsl1.aqueos.com
1399016431 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f
1399016542 J * opuk ~kupo@h-1-5.a176.priv.bahnhof.se
1399019850 Q * SteeleNivenson Read error: Connection reset by peer
1399022216 J * SteeleNivenson ~SteeleNiv@pool-108-29-139-222.nycmny.fios.verizon.net
1399022624 Q * Defaultti Quit: Quitting.
1399022742 J * Defaultti defaultti@lakka.kapsi.fi
1399031444 N * Bertl_zZ Bertl
1399031467 M * Bertl morning folks!
1399031596 N * l0kit Guest8388
1399031602 J * l0kit ~1oxT@0001b54e.user.oftc.net
1399031969 Q * Guest8388 Ping timeout: 480 seconds
1399032625 Q * SteeleNivenson Ping timeout: 480 seconds
1399032818 Q * Aiken Remote host closed the connection
1399038154 J * beng_ ~BenG@cpc29-aztw22-2-0-cust128.18-1.cable.virginm.net
1399041926 Q * Guy- Quit: electrician++
1399043975 Q * FireEgl Remote host closed the connection
1399045000 J * FireEgl ~FireEgl@173-23-77-153.client.mchsi.com
1399046011 J * Guy- ~korn@elan.rulez.org
1399047297 Q * beng_ Quit: I Leave
1399050820 J * SteeleNivenson ~SteeleNiv@pool-108-29-139-222.nycmny.fios.verizon.net
1399058155 Q * Ghislain Quit: Leaving.
1399059053 M * Guy- I upgraded a box from 3.7.7-vs2.3.5.6 to 3.13.11+vs2.3.6.11 and now udev doesn't create device nodes anymore... I wonder why?
1399059124 M * Bertl because of a change in ext4 it seems
1399059144 M * Bertl (which clashes with the mnt path checks done in Linux-VServer)
1399059238 J * bonbons ~bonbons@2001:a18:20c:401:f584:b0cf:54b8:ee00
1399059513 M * Guy- Bertl: oh.
1399059527 M * Guy- Bertl: do you know what the latest kernel that's expected to work is?
1399059558 M * Bertl IIRC the change came with 3.7.10/11
1399059575 M * Bertl there is a fixup patch on the ML
1399059646 M * Guy- that's susprising, because I have a box with 3.13.8-vs2.3.6.11 that doesn't have this problem
1399059700 M * Guy- (no ext4 filesystems need to be mounted for the problem to manifest, fwiw)
1399059743 M * distemper i guess the change came with the stable patch series (into all current stable kernels)...
1399059756 M * distemper 3.13.9 is broken e.g.
1399059844 M * Bertl Guy-: really, what filesystem is your rootfs?
1399059848 M * Guy- Bertl: xfs
1399059938 M * Bertl okay, thanks for the info
1399059978 M * Guy- do you have a link for the fixup patch?
1399059986 M * Guy- I'm not subscribed to the list
1399060003 M * Guy- (or I guess I can build a 3.13.8 kernel for thix box too)
1399060080 M * Bertl http://blog.gmane.org/gmane.linux.vserver
1399060094 M * distemper http://list.linux-vserver.org/archive?mss:6643:201404:adkhnakgaeglpfpaaakd
1399060106 M * Bertl ah, even better, thanks!
1399060128 M * Guy- just found it myself
1399060152 M * Guy- are you certain this is the same issue?
1399060162 M * Guy- my problem is that not even /dev/null and /dev/console get created
1399060188 M * Bertl very likely
1399060194 M * distemper yes, i had the same problem... my udev is not crashing, but the devtmpfs is empty
1399060272 M * Guy- okay, I'll try to revert this hunk and build a new kernel, then see if that helps
1399060285 M * distemper newer udev version don't create any devices anymore, devtmpfs comes pre-populated now
1399060348 M * Guy- yes, I gathered that much meanwhile
1399060358 M * Guy- but is it also devtmpfs that creates nodes later on?
1399060375 M * Guy- like e.g. when lvm is activated
1399060397 Q * bonbons Quit: Leaving
1399060432 M * Bertl I remember the lenghty discussion on lkml why a kernel created devfs is a very bad thing ... which was the reason for udev and switching to a tmpfs lateron (with a lot of hikkups :)
1399060433 M * distemper i think so, but you can still create them from userspace, of cause
1399060461 M * distemper yeah and that's why devtmpfs was also called "the return of devfs" ;)
1399060465 M * Bertl now, a few years later, we have a brand new development: devtmpfs :)
1399060480 M * Guy- yes, I also remember devfs
1399060483 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f
1399060496 M * Guy- for all the alleged unfixable theoretical problems, it worked pretty well
1399060546 M * Guy- distemper: to be able to boot I created an init-top/mknod script that created the device nodes :)
1399060571 M * Guy- but of course that's not a solution
1399060611 M * Bertl sounds like a perfectly fine userspace solution :)
1399060613 M * Guy- installing the new kernel on this box takes longer than compiling it on the other one
1399060617 J * bonbons ~bonbons@2001:a18:20c:401:f584:b0cf:54b8:ee00
1399060647 M * Guy- Bertl: to enable booting, yes :) but I had all sorts of problems later, for example I couldn't ssh into the box because the pty devices were missing
1399060649 M * daniel_hozac lol, what kind of box is it?
1399060669 M * Guy- daniel_hozac: the compile took 58s :)
1399060714 M * Guy- where I'm installing it is an old phenom II X3 705e box
1399060730 M * Guy- with unusually slow i/o for some reason (I haven't yet found out why)
1399060800 M * Guy- also, when I install a kernel it builds the spl+zfs modules using dkms
1399060803 M * Guy- and that's slow
1399060854 M * Guy- ah, the i/o slowness seems to be due to a WD WD10EADS drive
1399060907 M * daniel_hozac ah
1399061022 J * SteeleNivenson_ ~SteeleNiv@pool-108-29-139-222.nycmny.fios.verizon.net
1399061247 M * Guy- still installing...
1399061286 M * distemper Bertl: it's this commit https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=b37199e626b31e1175fb06764c5d1d687723aac2 btw (maybe the commit message helps or sth)
1399061470 M * Guy- aaaaaand reboot
1399061480 M * Bertl distemper: ah, yeah, thanks
1399061589 M * Guy- you were right, it was this problem!
1399064241 M * Guy- # mount vserver /dev/cgroup -t cgroup -o hugetlb,blkio,freezer,devices,memory,cpuacct,cpu                                                          :(
1399064245 M * Guy- mount: special device vserver does not exist
1399064247 M * Guy- wtf?
1399064320 M * daniel_hozac what does strace say? i'm guessing it's mount's smarts failing.
1399064327 M * Guy- mount("vserver", "/dev/cgroup", "cgroup", MS_MGC_VAL, "hugetlb,blkio,freezer,devices,memory,cpuacct,cpu") = -1 ENOENT (No such file or directory)
1399064339 M * daniel_hozac does /dev/cgroup exist?
1399064352 M * Guy- # CONFIG_CGROUP_HUGETLB is not set
1399064355 M * Guy- this might be it
1399064365 M * daniel_hozac perhaps
1399064367 M * Guy- yes, that was the problem
1399065893 J * vasko ~vasko@unreal.rainside.sk
1399071896 J * zerick ~eocrospom@190.187.21.53
1399072482 Q * bonbons Quit: Leaving