[mythtv-users] Same error every day ~5:30 pm

[George Bingham]

On Sat, Nov 5, 2022 at 1:50 PM Klaas de Waal <klaas.de.waal at gmail.com>
wrote:

> Hi George, all,
>
> For testing I have added an explicit QList out-of-range addressing error
> just to see what happens on my Fedora system.
> When configured with "--compile-type=release" the invalid QList access
> gives some number back.
> When configured with "--compile-type=debug" or without specifying a
> --compile-type option the invalid QList access fails with the same ASSERT
> failure "index out of range".
> The actual Qt error is given by the Q_ASSERT_X macro and this macro is
> documented as follows:
> Q_ASSERT_X is useful for testing pre- and post-conditions during
> development. It does nothing if QT_NO_DEBUG was defined during
> compilation.
> Apparently the mythtv packages for Ubuntu are not compiled for release.
>
> I have tried to reproduce the problem by setting up two recording rules
> that keep one recording, but this does not give the ASSERT failure.
> Possibly the ASSERT failure does depend on some other setting in the
> recording rule?
> Can you give some more details about how the recording rule is configured?
>
> If the problem cannot be reproduced by me then I can only have a look at a
> gdb traceback that you provide.
> Attaching gdb to an already started mythbackend as suggested by Mike,
> before the crucial time of 5:30, sounds like a plan.
>
> Thanks,
> Klaas.
>
>
> On Wed, 2 Nov 2022 at 20:08, Mike Hodson <mystica at gmail.com> wrote:
>
>>
>>
>> On Wed, Nov 2, 2022, 12:40 George Bingham <georgeb1962 at gmail.com> wrote:
>>
>>>
>>>
>>> On Wed, Nov 2, 2022 at 5:18 AM Klaas de Waal <klaas.de.waal at gmail.com>
>>> wrote:
>>>
>>>>
>>>>
>>>>
>>>> It would be great if you can start mythbackend from gdb and then, when
>>>> the segfault has happened, create a backtrace like this:
>>>> gdb> bt full
>>>> and attach the output to the ticket.
>>>>
>>>>
>>>
>>> I am not sure how to run the backend through gdb. Not sure I even have
>>> gdb on my system unless it's there by default. (OK, just checked, it is)
>>> The backend is started by the system, if there's a simple change I can make
>>> to have it launch through gdb automatically I can do that for the next few
>>> days, but then I'd need to know how to deal with the output... ??
>>>
>>
>> What you could do is attach gdb to the running process after it launches,
>> so you don't have to do anything special to the boot up process.
>>
>>
>> https://stackoverflow.com/questions/2308653/can-i-use-gdb-to-debug-a-running-process
>> has a couple good discussions of different ways to do it, from command line
>> options to launching gdb in interactive mode and using commands within it.
>>
>> You would want to run GDB as the user that is running the myth TV
>> process, so you might need a sudo -u mythtv prefixing the gdb command, or
>> launch a shell as the mythtv user and interact with gdb that way.
>>
>> That said, I do not know if there are any apparmor permissions or such
>> that might need to be added for the mythtv user to debug its own process,
>> so there is also the slight chance that the easiest way to do this,
>> although theoretically not the safest or most secure, would be to run it as
>> root.
>>
>> Hope this helps!
>>
>> Mike
>>
>> _______________________________________________
>> mythtv-users mailing list
>> mythtv-users at mythtv.org
>> http://lists.mythtv.org/mailman/listinfo/mythtv-users
>> http://wiki.mythtv.org/Mailing_List_etiquette
>> MythTV Forums: https://forum.mythtv.org
>>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://lists.mythtv.org/mailman/listinfo/mythtv-users
> http://wiki.mythtv.org/Mailing_List_etiquette
> MythTV Forums: https://forum.mythtv.org



Thanks Klaas,

It didn't happen on Thursday, which was OK because I wasn't home yet and
wouldn't have been able to trace it. Nor did it happen at 5:30 on Friday
even though I was ready for it...

It did happen sometime between 7: and 7:30 last night, but I wasn't
expecting it so was not ready. I''m not sure what caused it last night
because it wasn't at the point of starting a recording, rather it was in
the middle of two recordings.

One of my problems is that my myth backend drives all four of our TVs, and
we each have unique schedules which keep it busy almost 24 / 7... I'm being
a bit over dramatic, but (not me) 2 members of my house like to leave the
tv on overnight while sleeping, and the android frontends do not recover
from this without intervention.

There is a new frontend / backend waiting in my updates, and later today we
will be streaming our college football games that don't come over the air,
so I can update while that is happening and I will attach to it then and
wait and see if it happens this evening.

-- George
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mythtv.org/pipermail/mythtv-users/attachments/20221105/724ff9f2/attachment.htm>