linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Nikolay Borisov <nborisov@suse.com>
To: "Scott E. Blomquist" <sb@techsquare.com>,
	Jojo <jojo@automatix.de>,
	linux-btrfs@vger.kernel.org
Subject: Re: btrfs hang on nfs?
Date: Thu, 10 Jan 2019 13:51:43 +0200	[thread overview]
Message-ID: <f95b63e3-9da5-ea60-3b4c-fbdf223f439c@suse.com> (raw)
In-Reply-To: <23607.12444.740949.683554@techsquare.com>



On 10.01.19 г. 13:46 ч., Scott E. Blomquist wrote:
> 
> Scott E. Blomquist writes:
>  > 
>  > Thank you so much, I would have never guessed that.
>  > 
>  > In my case the scheduler was set to deadline
>  > 
>  >     cat /sys/block/sd[a-c]/queue/scheduler
>  >     noop [deadline] cfq
>  >     noop [deadline] cfq
>  >     noop [deadline] cfq
>  > 
>  > I have set it to cfq, I did not see an option for bfq.
>  > 
>  > I am also building 4.19.13 and will swap that in ASAP.
>  > 
> 
> Unfortunately the cfq scheduler did not help.  The system wedged.
> 
> I did notice this for the first time...
> 
>     [Wed Jan  9 06:03:41 2019] BTRFS info (device sda1): the free space cache file (83320273633280) is invalid, skip it

What you could do is mount btrfs with -o clear_cache to make btrfs
rebuild the freespace cache.

> 
> anything I should do about that?
> 
> The messages were similar...
> 
>     [Wed Jan  9 23:52:04 2019] INFO: task nfsd:2997 blocked for more than 120 seconds.
>     [Wed Jan  9 23:52:04 2019]       Not tainted 4.17.14-custom #1
>     [Wed Jan  9 23:52:04 2019] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
>     [Wed Jan  9 23:52:04 2019] nfsd            D    0  2997      2 0x80000000
>     [Wed Jan  9 23:52:04 2019] Call Trace:
>     [Wed Jan  9 23:52:04 2019]  ? __schedule+0x2cf/0x850
>     [Wed Jan  9 23:52:04 2019]  ? iput+0x6f/0x1b0
>     [Wed Jan  9 23:52:04 2019]  schedule+0x32/0x80
>     [Wed Jan  9 23:52:04 2019]  rwsem_down_write_failed+0x1e0/0x350
>     [Wed Jan  9 23:52:04 2019]  call_rwsem_down_write_failed+0x13/0x20
>     [Wed Jan  9 23:52:04 2019]  down_write+0x29/0x40
>     [Wed Jan  9 23:52:04 2019]  btrfs_file_write_iter+0xac/0x570 [btrfs]
>     [Wed Jan  9 23:52:04 2019]  ? nfsd_setuser+0x103/0x270 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  do_iter_readv_writev+0xff/0x150
>     [Wed Jan  9 23:52:04 2019]  do_iter_write+0x78/0x180
>     [Wed Jan  9 23:52:04 2019]  nfsd_vfs_write+0xf0/0x440 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  nfsd_write+0x84/0x150 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  nfsd3_proc_write+0xcc/0x150 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  nfsd_dispatch+0xb7/0x250 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  svc_process_common+0x382/0x730 [sunrpc]
>     [Wed Jan  9 23:52:04 2019]  svc_process+0xeb/0x100 [sunrpc]
>     [Wed Jan  9 23:52:04 2019]  nfsd+0xe3/0x150 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  kthread+0xf8/0x130
>     [Wed Jan  9 23:52:04 2019]  ? nfsd_destroy+0x60/0x60 [nfsd]
>     [Wed Jan  9 23:52:04 2019]  ? kthread_bind+0x10/0x10
>     [Wed Jan  9 23:52:04 2019]  ret_from_fork+0x35/0x40
> ... these two were different and at the tail of the dmesg...
>     [Wed Jan  9 23:52:05 2019] INFO: task kworker/u18:8:24049 blocked for more than 120 seconds.
>     [Wed Jan  9 23:52:05 2019]       Not tainted 4.17.14-custom #1
>     [Wed Jan  9 23:52:05 2019] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
>     [Wed Jan  9 23:52:05 2019] kworker/u18:8   D    0 24049      2 0x80000000
>     [Wed Jan  9 23:52:05 2019] Workqueue: writeback wb_workfn (flush-btrfs-1)
>     [Wed Jan  9 23:52:05 2019] Call Trace:
>     [Wed Jan  9 23:52:05 2019]  ? __schedule+0x2cf/0x850
>     [Wed Jan  9 23:52:05 2019]  schedule+0x32/0x80
>     [Wed Jan  9 23:52:05 2019]  wait_current_trans+0x87/0xd0 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  ? wait_woken+0x80/0x80
>     [Wed Jan  9 23:52:05 2019]  start_transaction+0x1ef/0x3f0 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  find_free_extent+0x104b/0x1060 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  ? alloc_debug_processing+0x7a/0x12c
>     [Wed Jan  9 23:52:05 2019]  btrfs_reserve_extent+0xf6/0x1f0 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  cow_file_range.isra.67+0x119/0x440 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  run_delalloc_range+0xff/0x380 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  writepage_delalloc+0xfb/0x150 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  __extent_writepage+0xdf/0x2b0 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  ? invalid_page_referenced_vma+0x90/0x90
>     [Wed Jan  9 23:52:05 2019]  extent_write_cache_pages+0x22f/0x3d0 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  extent_writepages+0x3f/0x60 [btrfs]
>     [Wed Jan  9 23:52:05 2019]  do_writepages+0x1a/0x70
>     [Wed Jan  9 23:52:05 2019]  __writeback_single_inode+0x3d/0x340
>     [Wed Jan  9 23:52:05 2019]  writeback_sb_inodes+0x224/0x480
>     [Wed Jan  9 23:52:05 2019]  __writeback_inodes_wb+0x87/0xb0
>     [Wed Jan  9 23:52:05 2019]  wb_writeback+0x258/0x2f0
>     [Wed Jan  9 23:52:05 2019]  ? cpumask_next+0x16/0x20
>     [Wed Jan  9 23:52:05 2019]  wb_workfn+0x2df/0x430
>     [Wed Jan  9 23:52:05 2019]  process_one_work+0x14c/0x350
>     [Wed Jan  9 23:52:05 2019]  worker_thread+0x47/0x3e0
>     [Wed Jan  9 23:52:05 2019]  kthread+0xf8/0x130
>     [Wed Jan  9 23:52:05 2019]  ? rescuer_thread+0x2f0/0x2f0
>     [Wed Jan  9 23:52:05 2019]  ? kthread_bind+0x10/0x10
>     [Wed Jan  9 23:52:05 2019]  ret_from_fork+0x35/0x40
>     [Wed Jan  9 23:54:07 2019] INFO: task dsmc:17747 blocked for more than 120 seconds.
>     [Wed Jan  9 23:54:07 2019]       Not tainted 4.17.14-custom #1
>     [Wed Jan  9 23:54:07 2019] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
>     [Wed Jan  9 23:54:07 2019] dsmc            D    0 17747      1 0x00000000
>     [Wed Jan  9 23:54:07 2019] Call Trace:
>     [Wed Jan  9 23:54:07 2019]  ? __schedule+0x2cf/0x850
>     [Wed Jan  9 23:54:07 2019]  ? btrfs_set_path_blocking+0x31/0x60 [btrfs]
>     [Wed Jan  9 23:54:07 2019]  schedule+0x32/0x80
>     [Wed Jan  9 23:54:07 2019]  wait_current_trans+0x87/0xd0 [btrfs]
>     [Wed Jan  9 23:54:07 2019]  ? wait_woken+0x80/0x80
>     [Wed Jan  9 23:54:07 2019]  start_transaction+0x1ef/0x3f0 [btrfs]
>     [Wed Jan  9 23:54:07 2019]  btrfs_dirty_inode+0x49/0xe0 [btrfs]
>     [Wed Jan  9 23:54:07 2019]  touch_atime+0x8c/0xb0
>     [Wed Jan  9 23:54:07 2019]  iterate_dir+0x10a/0x180
>     [Wed Jan  9 23:54:07 2019]  __x64_sys_getdents+0x8e/0x110
>     [Wed Jan  9 23:54:07 2019]  ? iterate_dir+0x180/0x180
>     [Wed Jan  9 23:54:07 2019]  do_syscall_64+0x55/0x110
>     [Wed Jan  9 23:54:07 2019]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>     [Wed Jan  9 23:54:07 2019] RIP: 0033:0x7f0c982f9215
>     [Wed Jan  9 23:54:07 2019] RSP: 002b:00007f0c918a95c0 EFLAGS: 00000246 ORIG_RAX: 000000000000004e
>     [Wed Jan  9 23:54:07 2019] RAX: ffffffffffffffda RBX: 00007f0c3f348c60 RCX: 00007f0c982f9215
>     [Wed Jan  9 23:54:07 2019] RDX: 0000000000008000 RSI: 00007f0c3f348c60 RDI: 0000000000000007
>     [Wed Jan  9 23:54:07 2019] RBP: 00007f0c3f348c60 R08: 00000000000000ff R09: 0000000000001000
>     [Wed Jan  9 23:54:07 2019] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000a6672f
>     [Wed Jan  9 23:54:07 2019] R13: 00007f0c918affe0 R14: 0000000000000000 R15: 00007f0c918affb0
> 

These don't tell the full story, what seems to be happening is that
stuff is waiting for transaction to finish but it's not evident which
thread is holding the transaction. Please, paste the output of
"echo w > /proc/sysrq-trigger" so we have full picture of what's blocked
where.


> 

  reply	other threads:[~2019-01-10 11:51 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-01-09 10:54 btrfs hang on nfs? Scott E. Blomquist
2019-01-09 12:14 ` Jojo
2019-01-09 13:31   ` Scott E. Blomquist
2019-01-10 11:46     ` Scott E. Blomquist
2019-01-10 11:51       ` Nikolay Borisov [this message]
2019-01-10 12:00         ` Scott E. Blomquist
2019-01-14 11:42           ` Scott E. Blomquist
2019-01-14 12:11             ` Nikolay Borisov
2019-01-14 13:13               ` Scott E. Blomquist
2019-01-14 13:28                 ` Nikolay Borisov
2019-01-15  8:16                 ` Nikolay Borisov
2019-01-15 14:36                   ` Scott E. Blomquist

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f95b63e3-9da5-ea60-3b4c-fbdf223f439c@suse.com \
    --to=nborisov@suse.com \
    --cc=jojo@automatix.de \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=sb@techsquare.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).