On Fri, May 18, 2018 at 10:14 AM, Yan, Zheng <zyan@xxxxxxxxxx> wrote:
>
>> On May 17, 2018, at 01:37, Ilya Dryomov <idryomov@xxxxxxxxx> wrote:
>>
>> On Mon, May 14, 2018 at 11:51 AM, Yan, Zheng <zyan@xxxxxxxxxx> wrote:
>>>
>>> [...]
>>>
>>> Sorry, I was wrong about where he hang occurs. It’s at
>>>
>>> [<0>] io_schedule+0xd/0x30
>>> [<0>] wait_on_page_bit_common+0xc6/0x130
>>> [<0>] __filemap_fdatawait_range+0xbd/0x100
>>> [<0>] filemap_fdatawait_keep_errors+0x15/0x40
>>> [<0>] sync_inodes_sb+0x1cf/0x240
>>> [<0>] sync_filesystem+0x52/0x90
>>> [<0>] generic_shutdown_super+0x1d/0x110
>>> [<0>] ceph_kill_sb+0x28/0x80 [ceph]
>>> [<0>] deactivate_locked_super+0x35/0x60
>>> [<0>] cleanup_mnt+0x36/0x70
>>> [<0>] task_work_run+0x79/0xa0
>>> [<0>] exit_to_usermode_loop+0x62/0x70
>>> [<0>] do_syscall_64+0xdb/0xf0
>>> [<0>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>> [<0>] 0xffffffffffffffff
>>
>> Makes sense, please make sure to update the commit message.
>>
>> I pushed wip-umount-force and gave it a quick test. The filesystem
>> part appears to need improvement -- lots of "VFS: Busy inodes after
>> unmount of ceph", more often than not with various crashes. See the
>> attached.
>>
>
> I push a new commit to wip-umount-force, could you have a try.
Seems to help -- my (very simple) test case now passes.
Could you repush wip-umount-force with an updated commit message for
"ceph: abort osd requests on force umount" and I'll add it to testing?
Thanks,
Ilya
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
