Actual result: The machine hangs or freezes at the end of the snapshot.
You will notice the next message in the Linux logs:
Jan 10 00:12:55 x3421 kernel: INFO: task xfsaild/md125:827 blocked for more than 120 seconds.
Jan 10 00:12:55 x3421 kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Jan 10 00:12:55 x3421 kernel: xfsaild/md125 D ffff9b4dfed1ac80 0 827 2 0x00000000
Jan 10 00:12:55 x3421 kernel: Call Trace:
Jan 10 00:12:55 x3421 kernel: [] ? try_to_del_timer_sync+0x5e/0x90
Jan 10 00:12:55 x3421 kernel: [] schedule+0x29/0x70
Jan 10 00:12:55 x3421 kernel: [] _xfs_log_force+0x1c6/0x2a0 [xfs]
Jan 10 00:12:55 x3421 kernel: [] ? wake_up_state+0x20/0x20
Jan 10 00:12:55 x3421 kernel: [] ? xfsaild+0x180/0x760 [xfs]
Jan 10 00:12:55 x3421 kernel: [] xfs_log_force+0x2c/0x70 [xfs]
Jan 10 00:12:55 x3421 kernel: [] ? xfs_trans_ail_cursor_first+0x90/0x90 [xfs]
Jan 10 00:12:55 x3421 kernel: [] xfsaild+0x180/0x760 [xfs]
Jan 10 00:12:55 x3421 kernel: [] ? xfs_trans_ail_cursor_first+0x90/0x90 [xfs]
Jan 10 00:12:55 x3421 kernel: [] kthread+0xd1/0xe0
Jan 10 00:12:55 x3421 kernel: [] ? insert_kthread_work+0x40/0x40
Jan 10 00:12:55 x3421 kernel: [] ret_from_fork_nospec_begin+0xe/0x21
Jan 10 00:12:55 x3421 kernel: [] ? insert_kthread_work+0x40/0x40
Update the XFS drivers. If this doesn't help, schedule the snapshots to run at a time when there are fewer I/O operations.
Note: Yet that the issue is with the driver and not with the kernel we assume that it can be reproduced on a different OS.
