fsync: giving up on dirty

[Johannes Totz]

On 02/03/2012 21:36, Kirk McKusick wrote:
>> From: Ivan Voras <ivoras at freebsd.org>
>> Date: Fri, 2 Mar 2012 12:57:00 +0100
>> Subject: Re: fsync: giving up on dirty
>> To: Kirk McKusick <mckusick at mckusick.com>
>> Cc: freebsd-fs at freebsd.org
>>
>> On 29 February 2012 20:19, Kirk McKusick <mckusick at mckusick.com> wrote:
>>
>>> I believe that the problem is because the soft updates worklist needs
>>> to be flushed before some of the dirty blocks can be successfully written.
>>> If you are running a 9-stable system on this machine, are using journaled
>>> soft updates on the filesystem in question, and are willing to try out
>>> my first attempt at a fix, let me know and I'll send you the diffs for it.
>>
>> The thing is - I'm not. This is a 9-stable, but it was upgraded from 8
>> and I don't have SUJ enabled.
>>
>> I keep getting such messages daily.
>>
>> Mar  1 04:02:09 skynet kernel: fsync: giving up on dirty
>> Mar  1 04:02:09 skynet kernel: 0xfffffe000fef2780: tag devfs, type VCHR
>> Mar  1 04:02:09 skynet kernel: usecount 1, writecount 0, refcount 2515
>> mountedhere 0xfffffe000fd25200
>> Mar  1 04:02:09 skynet kernel: flags ()
>> Mar  1 04:02:09 skynet kernel: v_object 0xfffffe000fe96d98 ref 0 pages 10182
>> Mar  1 04:02:09 skynet kernel: lock type devfs: EXCL by thread
>> 0xfffffe003dc8d000 (pid 79344)
>> Mar  1 04:02:09 skynet kernel: dev multipath/hpdisk4-web
> 
> Thanks for the report. It is useful to know that this can occur even
> with non SUJ systems. I am working with Peter Holm to identify the
> cause and will keep you (and the list) posted with what we figure out.

I got one of these as well:

fsync: giving up on dirty
0xfffffe000d00bd20: tag devfs, type VCHR
    usecount 1, writecount 0, refcount 638 mountedhere 0xfffffe0003d23600
    flags (VV_COPYONWRITE)
    v_object 0xfffffe0003e44ca8 ref 0 pages 2538
    lock type devfs: EXCL by thread 0xfffffe001c075460 (pid 11645)
       dev ufs/var

This happened during snapshot creation for backup with dump(8).
So far machine is running fine. Subsequent snapshots succeeded.

This is on:

FreeBSD XXX 9.0-PRERELEASE FreeBSD 9.0-PRERELEASE #1 r227793M: Mon Dec
5 22:57:54 GMT 2011     root at XXX:/usr/obj/usr/src/sys/GENERIC  amd64

No journaling, just plain old UFS 2 with softupdates. Plenty of free
space available.



Johannes