On Wed, Oct 18, 2017 at 03:26:25PM +0800, Eryu Guan wrote:
> Hi Josef,
>
> On Tue, Oct 10, 2017 at 03:58:31PM -0400, Josef Bacik wrote:
> > From: Josef Bacik <jbacik@xxxxxx>
> >
> > This uses the new fio results perf helpers to run a rand write buffered
> > workload on the scratch device.
> >
> > Signed-off-by: Josef Bacik <jbacik@xxxxxx>
> > ---
> > v1->v2:
> > - updated to use the new _require_fio_results helper and moved the
> > _fio_results_init call to after teh _require_fio check
>
> I tried this v2 a bit, but met some problems, I haven't looked into the
> code closely though, just wanted to get a first impression of this perf
> test frame work.
>
> - missing perf/group file, so test won't be run by check
Sigh shit, I knew that was going to happen. Forgot to add it to the commit,
I'll add it in there.
> - this test writes 16G files to SCRATCH_DEV by default, and my device
> has only 15G, so fio failed with ENOSPC, I think we need a require
> rule on the test device
Oops yup I'll fix that.
> - after working around the group file and device size issue, test still
> failed like
>
> perf/001.full:
> ....
> # /usr/local/bin/fio --output-format=json --output=/tmp/30750.json /tmp/30750.fio
> # /usr/bin/python2 /root/xfstests/src/perf/fio-insert-and-compare.py -c default -d /root/xfstests/results//fio-results.db -n 001 /tmp/30750.json
> Traceback (most recent call last):
> File "/root/xfstests/src/perf/fio-insert-and-compare.py", line 28, in <module>
> result_data.insert_result(data)
> File "/root/xfstests/src/perf/ResultData.py", line 43, in insert_result
> self._insert_obj('fio_jobs', job)
> File "/root/xfstests/src/perf/ResultData.py", line 35, in _insert_obj
> cur.execute(cmd, tuple(values))
> sqlite3.IntegrityError: fio_jobs.trim_lat_ns_mean may not be NULL
> failed: '/usr/bin/python2 /root/xfstests/src/perf/fio-insert-and-compare.py -c default -d /root/xfstests/results//fio-results.db -n 001 /tmp/30750.json'
>
> Am I missing something? BTW, I was using fio-2.6-2.el7.x86_64.
Nope again I forgot to commit local changes, sorry about that. I'll fix this up
and try again. Thanks,
Josef
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
