You are not logged in.
Lately (at least after upgrading to linux 3.15), when I try to copy lots of files between btrfs partitions, I'm getting these "IO hangs" (not sure what to call it) where I can use the system fine as long as I don't try to read/write from my btrfs /home. So I can log in as root and look around, but as soon as I try and ls some random dir within my btrfs /home, it blocks and I have to log in from a different tty to use the system.
A "sudo journalctl -b-1 |grep -i btrfs" will show stuff like
juli 17 14:25:54 arch kernel: [<ffffffffa049fadf>] wait_current_trans.isra.18+0xcf/0x120 [btrfs]
juli 17 14:25:54 arch kernel: [<ffffffffa04a12d8>] start_transaction+0x3a8/0x5c0 [btrfs]
juli 17 14:25:54 arch kernel: [<ffffffffa04a1547>] btrfs_join_transaction+0x17/0x20 [btrfs]
juli 17 14:25:54 arch kernel: [<ffffffffa04a6dd3>] btrfs_dirty_inode+0x43/0xf0 [btrfs]
juli 17 14:25:54 arch kernel: [<ffffffffa04a6eff>] btrfs_update_time+0x7f/0xc0 [btrfs]
juli 17 14:35:54 arch kernel: INFO: task btrfs-transacti:282 blocked for more than 120 seconds.
juli 17 14:35:54 arch kernel: btrfs-transacti D 0000000000000000 0 282 2 0x00000000
juli 17 14:35:54 arch kernel: [<ffffffffa04b0df6>] btrfs_wait_and_free_delalloc_work+0x16/0x30 [btrfs]
juli 17 14:35:54 arch kernel: [<ffffffffa04bb128>] btrfs_run_ordered_operations+0x1f8/0x2e0 [btrfs]
juli 17 14:35:54 arch kernel: [<ffffffffa04a0546>] btrfs_commit_transaction+0x36/0xa20 [btrfs]
juli 17 14:35:54 arch kernel: [<ffffffffa049c475>] transaction_kthread+0x1e5/0x250 [btrfs]
juli 17 14:35:54 arch kernel: [<ffffffffa049c290>] ? btrfs_cleanup_transaction+0x5a0/0x5a0 [btrfs]
juli 17 14:35:54 arch kernel: Workqueue: btrfs-flush_delalloc normal_work_helper [btrfs](see full paste at http://pastie.org/9399449)
There seems to be others with similar issues:
http://www.spinics.net/lists/linux-btrfs/msg34458.html
http://www.mail-archive.com/linux-btrfs … 33795.html
but no solution from what I can tell.
It may be that it would unhang itself if I waited a really long while, but haven't had the patience to wait and see yet. I do run nightly backups between the btrfs partitions and the computer seems fine each morning though …
Anyone know what to do about this?
Last edited by unhammer (2014-07-17 13:49:55)
Offline
I'm seeing the same deadlock when trying to deal with a failed drive in a RAID10. The result being that I'm unable work with the degraded volume at all.
Offline
Same thing for me when running rsync... It also corrupted my filesystem. So I restored from a backup and I'm using XFS@LVM@dm-crypt for the time being.
https://imgur.com/a/6gNhH
Offline