[mythtv-users] Lots of mythpreviewgen processes left running on MBE

Tom Lichti redpepperracing at gmail.com
Sat Sep 18 19:05:19 UTC 2010


On Sat, Sep 18, 2010 at 12:59 PM, Gavin Hurlbut <gjhurlbu at gmail.com> wrote:
> On Sat, Sep 18, 2010 at 5:58 AM, Tom Lichti <redpepperracing at gmail.com>
> wrote:
>>
>> (gdb) thread 2
>> [Switching to thread 2 (Thread 0x2ec9b70 (LWP 3157))]#1  0x02068536 in
>> __poll (fds=0xb4b01d88, nfds=1, timeout=-1) at
>> ../sysdeps/unix/sysv/linux/poll.c:87
>> 87        int result = INLINE_SYSCALL (poll, 3, CHECK_N (fds, nfds),
>> nfds, timeout);
>> (gdb) print *fds
>> $6 = {fd = 0, events = 1, revents = 0}
>> (gdb) print fds
>> $7 = (struct pollfd *) 0xb4b01d88
>> (gdb) print nfds
>> $8 = 1
>
> Hmmm, seems to be blocked waiting for input on stdin.  No clue why it would
> be trying to do that.
>
>>
>> What is the process id? Or am I missing something?
>
> The same pid you attached to in gdb.  Do "lsof -p pid" with that number
> after the -p.
>>
>>
>> <USN>uuid:1216ac5d-850f-411e-81d3-34172d45c05e::urn:schemas-upnp-org:device:MediaServer:1</USN>
>
> OK, that's what I thought. :)  The good news is:  I have a commit going in
> this morning that should keep this from happening.  The issue seems to be
> when it tries to destroy an HTTP Server thread that we don't even need, and
> failed startup (probably due to another similar process having the port
> open).  This is during the SSDP search (UPnP backend searching), and it's
> only even tried with the USN set in the config.xml.
> My change removes the HTTP server completely, so it will no longer be
> conflicting with each other, nor failing to be destroyed.  So after I have
> breakfast, then get a chance to commit the changes, this hopefully will be
> gone.

Excellent! I just saw the commit go through, I'll try and recompile
shortly and let you know how it goes.

Thanks!
Tom


More information about the mythtv-users mailing list