I got this problem since 3.8.5 + for-linus (from that time). Have just tried 3.8.6 + for-linus with "git merge -X theirs btrfs/for-linus" but still same problem. Going back to 3.7.4 + for-linus (from that time) doesn''t give me the problem. This is an production nfs server with 2x2TB raid1, so cant reboot it that often. Have seen this same problem on another system (also raid1) once, but rebooting helped, no problems since. Both systems use autodefrag, maybe that sometimes triggers it? I really would like to help, so i can stay on the latest kernels. What should i do? Ahmet
On Wed, April 10, 2013 at 09:58 (+0200), Ahmet Inan wrote:> I got this problem since 3.8.5 + for-linus (from that time). > Have just tried 3.8.6 + for-linus with "git merge -X theirs > btrfs/for-linus" but still same problem. > Going back to 3.7.4 + for-linus (from that time) doesn''t give me the problem.The stack you attached shows a function added in the snapshot aware defrag patches (commit 38c227d8), added in 3.8. The real problem, however, is not caused by that commit but by a tree mod log bug. I expect that fs/btrfs/ctree.c:1144 is this BUG_ON in your kernel from __tree_mod_log_rewind (my line numbers don''t match): 1138 switch (tm->op) { 1139 case MOD_LOG_KEY_REMOVE_WHILE_FREEING: 1140 BUG_ON(tm->slot < n); I''ve got a fix for that I''m currently testing, expect it on the list soon.> This is an production nfs server with 2x2TB raid1, so cant reboot it that often. > Have seen this same problem on another system (also raid1) once, but > rebooting helped, no problems since. > Both systems use autodefrag, maybe that sometimes triggers it? > I really would like to help, so i can stay on the latest kernels. > What should i do?For the meantime I recommend to not defrag your filesystem. As a general remark, please send your stack traces inline, not as attachment if possible. Thanks, -Jan -- 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
> The real problem, however, is not caused by that commit but by a tree mod log > bug. I expect that fs/btrfs/ctree.c:1144 is this BUG_ON in your kernel from > __tree_mod_log_rewind (my line numbers don''t match): > > 1138 switch (tm->op) { > 1139 case MOD_LOG_KEY_REMOVE_WHILE_FREEING: > 1140 BUG_ON(tm->slot < n); > > I''ve got a fix for that I''m currently testing, expect it on the list soon.Yea, thats the line.> For the meantime I recommend to not defrag your filesystem.Ok, already disabled and rebooted. Will come back to You if i still get that Problem.> As a general remark, please send your stack traces inline, not as attachment if > possible.Google mail always screws longer than 80 lines in plain text mode and non plain text is not what You want. Maybe i should go back to using mutt. Thank You, Ahmet -- 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