Re: xfsaild in D state seems to be blocking all other i/o sporadically

[Brian Foster <bfoster@redhat.com>]

On Wed, Apr 19, 2017 at 06:36:00PM +0200, Michael Weissenbacher wrote:
> On 19.04.2017 16:04, Carlos Maiolino wrote:
> > If I am not blind, the task currently owning the cpu isn't listed there, which I
> > suppose to be already running.
> Correct, the only process utilizing the CPU is xfsaild. It is also the
> only process doing I/O, and it looks like it's doing only reads. It
> (xfsaild) never shows up in dmesg as blocked.
> 

A snippet of tracepoint output from the fs might also be interesting
when you hit this state. E.g., 'trace-cmd start -e xfs:*' and save the
output of /sys/kernel/debug/tracing/trace_pipe to a file for a few
seconds or so.

> > Next time you hit it, could you please paste the stack trace of the task owning
> > the cpu? Which I suppose to be xfsaild according to your report.
> > 
> > A simple `cat /proc/<pid>/stack` is probably enough
> Sure, will do.

BTW, is the original /proc/meminfo output you posted from the system in
a normal state or when the problem occurs? If the former, could you
include it while the problem occurs as well?

Brian

> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html