On Wed, 2018-01-31 at 13:22 +1100, Dave Chinner wrote:
> 
> 
> Could be a disk that is slow, or could be many other
> things. More information required:
> 
> http://xfs.org/index.php/XFS_FAQ#Q:_What_information_should_I_include_when_reporting_a_problem.3F
> 
> And everything else is backed up behind it trying to allocate
> inodes. There could be many, many reasons for that, and that's why
> we need more information to begin to isolate the cause.
> 
> Cheers,
> 
> Dave.
> 



> # trace-cmd record -e xfs\*
> before the problem occurs, and once it has occurred, kill the trace-cmd with ctrl-C, and then run:
> 
> # trace-cmd report > trace_report.txt

https://dumps.sy24.ru/trace_report.txt.bz2 (860MB)

Ok, I collected needed information.
Cant you now look into in?