I'll see if I can get gdb setup like that.
Below output shows spec 176 was the last job that ran ok, so you could try if it reproduces with 177. However, the job number exhibiting the hang usually is different every time.
\Rob
On 08/03/2018, 14:29, "Sitsofe Wheeler" <sitsofe@xxxxxxxxx> wrote:
A backtrace would be a good starting point but we need your gdb to be
able to be able to produce that... If you've managed to successfully
attach gdb to fio running something like
set logging file fiobacktrace.txt
set logging on
thread apply all bt
set logging off
and copying the contents of fiobacktrace.txt to here would be helpful.
If possible also take a note of the exact job that failed and send it
over as presumably we can just keep rerunning that to produce the
problem.
On 8 March 2018 at 10:51, Rob Scheepens <rob.scheepens@xxxxxxxxxxx> wrote:
> Just tested using fio-3.5: the first attempt already hung, at:
>
> PS C:\Python27> python .\profiler_win_dk.py
> * Running job: 'config' (20 secs / spec)
> ** Executing spec 176/405 at 2018-03-08 02:10:25.360000
>
> Instead of trying an large amount of variations to reproduce, which will be time-consuming, is there a way to capture a dump using gdb and debug that? I have very limited gdb skills and when I recently tried to save it off, it gave me an error that seemed to implicate some lib was not implemented on Windows.
>
> I left the vm in the current state, in case anyone wants me to get something from it.
--
Sitsofe | https://urldefense.proofpoint.com/v2/url?u=http-3A__sucs.org_-7Esits_&d=DwIFaQ&c=s883GpUCOChKOHiocYtGcg&r=OMged-t_5I_fmfpUaT3vaA06lgLL_alYnDQJxHmXz64&m=_j-o6LW-O_G6OkO-vy7HdE7UJDsy8W_csQytzi2Vg7I&s=u17mNDv5miOlk6AgsMohHL41aIameCmzLB1vhzTlxJc&e=
��.n��������+%������w��{.n�������^n�r������&��z�ޗ�zf���h���~����������_��+v���)ߣ�
