On Sun, Jul 3, 2011 at 1:09 PM, Stephane Chazelas <stephane_chaze...@yahoo.fr> wrote: > Hiya, > > I've got a server using brtfs to implement a backup system. > Basically, every night, for a few machines, I rsync (and other > methods) their file systems into one btrfs subvolume each and > then snapshot it. > > On that server, the btrfs fs is on 3 3TB drives, mounted with > compress-force. Every week, I rsync the most recent snapshot of > a selection of subvolumes onto an encrypted (LUKS) external hard > drive (3TB as well). > > Now, on a few occasions (actually, most of the time), when I > rsynced the data (about 2.5TB) onto the external drive, the > system would crash after some time with "Out of memory and no > killable process". Basically, something in kernel was allocating > the whole memory, then oom mass killed everybody and crash. > > That was with ubuntu 2.6.38. I had then moved to debian and > 2.6.39 and thought the problem was fixed, but it just happened > again with 3.0.0rc5 while rsyncing onto an initially empty btrfs > fs. > > I'm going to resume the rsync again, and it's likely to happen > again. Is there anything simple (as I've got very little time to > look into that) I could do to help debug the issue (I'm not 100% > sure it's btrfs fault but that's the most likely culprit). > > For a start, I'll switch the console to serial, and watch > /proc/vmstat. Anything else I could do? > > Note that that server has never crashed when doing a lot of IO > at the same time in a lot of subvolumes with remote hosts. It's > only when copying data to that external drive on LUKS that it > seems to crash.
Look at the output of slabtop (should be installed by default, procfs package), before rsync for comparison, and during. -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html