[Question] how to avoid process hang without disk space

[Question] how to avoid process hang without disk space

[Gavin Shan]

Recently, we're encountering one problem that caused by full disk drives.
I'm not sure if there has any method to recover from the situation?

When the problem happened, I couldn't kill the process with signal 9.
Also, the "reboot" would cause system hang as well.

Here's the backtrace from kernel.
=================================

INFO: task jbd2/dm-0-8:463 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
jbd2/dm-0-8   D 0000000000000000     0   463      2 0x00008000
Call Trace:
[c0000006ef23f190] [c0000006ef23f240] 0xc0000006ef23f240 (unreliable)
[c0000006ef23f360] [c000000000014288] .__switch_to+0xf8/0x1d0
[c0000006ef23f3f0] [c00000000059cc18] .schedule+0x408/0xd30
[c0000006ef23f6d0] [c00000000059d5d0] .io_schedule+0x90/0x110
[c0000006ef23f760] [c00000000014c5a0] .sync_page+0x70/0xa0
[c0000006ef23f7e0] [c00000000059dee4] .__wait_on_bit_lock+0xd4/0x1b0
[c0000006ef23f8a0] [c00000000014c4e4] .__lock_page+0x54/0x70
[c0000006ef23f960] [c0000000001680bc] .write_cache_pages+0x3fc/0x4a0
[c0000006ef23fb10] [d000000004fb3384] .journal_submit_inode_data_buffers+0x64/0x90 [jbd2]
[c0000006ef23fc20] [d000000004fb3a1c] .jbd2_journal_commit_transaction+0x44c/0x1940 [jbd2]
[c0000006ef23fde0] [d000000004fbc2fc] .kjournald2+0xec/0x300 [jbd2]
[c0000006ef23fed0] [c0000000000bb02c] .kthread+0xbc/0xd0
[c0000006ef23ff90] [c0000000000323f4] .kernel_thread+0x54/0x70

Thanks,
Gavin

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>