Hello, ''space info full'' was the only message I got from dmesg. df reports: /dev/sdb1 33551720 28504548 5047172 85% /mnt/btrfstest df -i reports: /dev/sdb1 0 0 0 - /mnt/btrfstest Any (e.g., touch, rm) operation on that mounted filesystem lets the CPU rise and the system hang... I was filling up the partition with compilebench to the ration mentioned above. During this process I tried to create a subvolume with btrfsctl -s sub_vol /mnt/btrfstest which did not, as opposed to the advices from the INSTALL file, create anything within /mnt/btrfstest. Regards, -- Martin Bürger 1024D/27C9019B@www.edu.uni-klu.ac.at/~mbuerger/key.asc Key fingerprint = D10D 97EF 0C32 B337 5A12 0B6C 2D47 7575 27C9 019B -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Martin Bürger wrote:> Hello, > ''space info full'' was the only message I got from dmesg. df reports: > > /dev/sdb1 33551720 28504548 5047172 > 85% /mnt/btrfstestThe output of df can be inaccurate. Although df shows the FS is 85% full, It''s possible the FS is completely full.> > df -i reports: > /dev/sdb1 0 0 0 - /mnt/btrfstest > > Any (e.g., touch, rm) operation on that mounted filesystem lets the > CPU rise and the system hang...btrfs can''t properly handle disk full condition currently. ENOSPC can result in BUG()> > I was filling up the partition with compilebench to the ration > mentioned above. During this process I tried to create a subvolume > with > btrfsctl -s sub_vol /mnt/btrfstest > which did not, as opposed to the advices from the INSTALL file, > create anything within /mnt/btrfstest.Please try mount the FS with -o subvol=. option, you will see the new subvol. Regards Yan Zheng -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sat, Oct 04, 2008 at 09:45:55AM +0200, Martin Bürger wrote:> Hello, > ''space info full'' was the only message I got from dmesg. df reports: >space info full doesn''t mean the disk is full, just means we''ve allocated all of the block group chunks that the filesystem can hold, so you will get that message well before you actually run out of disk space. As it stands now btrfs has a short circuit when you hit ~15% free to go ahead an ENOSPC to avoid hitting the many BUG()''s that are sitting around to handle ENOSPC.> /dev/sdb1 33551720 28504548 5047172 > 85% /mnt/btrfstest > > df -i reports: > /dev/sdb1 0 0 0 - /mnt/btrfstest > > Any (e.g., touch, rm) operation on that mounted filesystem lets the > CPU rise and the system hang... >are you running the latest git trees of both btrfs-unstable and btrfs-progs-unstable? I''d expect BUG()''s or something like that, shouldn''t be getting hangs. Thanks, Josef -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Monday 06 October 2008, Josef Bacik wrote:> are you running the latest git trees of both btrfs-unstable and > btrfs-progs-unstable? I''d expect BUG()''s or something like that, > shouldn''t be getting hangs. Thanks,Yes, I used both unstable versions from the git tree for the tests. No further messages than the one I mentioned occurred. Furthermore the system did hang at any operation on that particular file system after (can not exactly tell when) the ''space info full'' message had turned up. Unfortunately I''ve not been able to reproduce this (yet), hence this might have been an issue with the virtual machine (VMWare) I using for the tests. Regards, -- Martin Bürger 1024D/27C9019B@www.edu.uni-klu.ac.at/~mbuerger/key.asc Key fingerprint = D10D 97EF 0C32 B337 5A12 0B6C 2D47 7575 27C9 019B -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html