[mythtv-users] Backend Response Problems

Wed Jan 27 10:21:34 UTC 2010

Hi -  please help! 

My backend will lock up - in the sense that the frontends either come to
a crawl and do not respond or display the message backend has gone away.
However *some* backend log entries are still being entered. 

Backend crashes didn't used to be a major problem as pcsk would restart
the process - now pcsk doesn't detect the problem and therefore doesnt
restart. 

I have a feeling its when I goto watch recordings that the problem
starts. However as this doesn't happen every time and cannot reproduce. 
I did see a post on the list about the lag associated with watch
recordings because of fan art. There was some instructions to comment
out a bit of text in a recordings-ui.xml  theme file. I tried this
hoping it would stop my crashing. Nope.

I set my logging onto 'most' with the hope of figuring out what was
wrong.  Last few backend log entries:

[Frontend became unresponsive at approx 21:52]
[2010-01-26 21:54:34] 2010-01-26 21:54:34.389 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:34] 2010-01-26 21:54:34.691 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:34] 2010-01-26 21:54:34.992 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:35] 2010-01-26 21:54:35.267 PESPacket: Failed CRC check 0xb6215433 != 0xc3bfa4b5 for StreamID = 0x70
[2010-01-26 21:54:35] 2010-01-26 21:54:35.301 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:35] 2010-01-26 21:54:35.603 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:35] 2010-01-26 21:54:35.905 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:36] 2010-01-26 21:54:36.207 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:36] 2010-01-26 21:54:36.275 PESPacket: Failed CRC check 0xb6215434 != 0xc3bfa4b5 for StreamID = 0x70
[2010-01-26 21:54:36] 2010-01-26 21:54:36.509 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)
[2010-01-26 21:54:36] 2010-01-26 21:54:36.812 SM(/dev/dvb/adapter0/frontend0)::AddFlags: Seen() Match() Wait(Sig,)

It also looked like EIT was happening close to the time of the failure.
However I cannot see any reason why it would be the cause of the
failure.

Another occurrence of this problem day before - here are the last few
entries of that log:

[2010-01-25 21:11:53] 2010-01-25 21:11:53.277 adding: droom as a client (events: 1)
[2010-01-25 21:11:53] 2010-01-25 21:11:53.277 MythSocket(ffffffff9863fe68:127): write -> 127 2       OK
[2010-01-25 21:11:53] 2010-01-25 21:11:53.277 MythSocket(ffffffff9863fe68:127): DownRef: 1
[2010-01-25 21:11:53] 2010-01-25 21:11:53.278 MythSocketThread(sock 0xffffffff9862eae0:125): socket is readable
[2010-01-25 21:11:53] 2010-01-25 21:11:53.278 MythSocketThread(sock 0xffffffff9862eae0:125): calling m_cb->readyRead()
[2010-01-25 21:11:53] 2010-01-25 21:11:53.278 MythSocket(ffffffff9862eae0:125): UpRef: 2
[2010-01-25 21:11:53] 2010-01-25 21:11:53.278 MythSocket(ffffffff9862eae0:125): read  <- 125 15      QUERY_TIME_ZONE
[2010-01-25 21:11:53] 2010-01-25 21:11:53.279 MythSocket(ffffffff9862eae0:125): write -> 125 43      Europe/London[]:[]0[]:[]2010-01-25T21:11:53
[2010-01-25 21:11:53] 2010-01-25 21:11:53.279 MythSocket(ffffffff9862eae0:125): DownRef: 1
[2010-01-25 21:11:54] 2010-01-25 21:11:54.850 MythSocketThread(sock 0xffffffff9862eae0:125): socket is readable
[2010-01-25 21:11:54] 2010-01-25 21:11:54.850 MythSocketThread(sock 0xffffffff9862eae0:125): calling m_cb->readyRead()
[2010-01-25 21:11:54] 2010-01-25 21:11:54.850 MythSocket(ffffffff9862eae0:125): UpRef: 2
[2010-01-25 21:11:54] 2010-01-25 21:11:54.850 MythSocket(ffffffff9862eae0:125): read  <- 125 33      QUERY_IS_ACTIVE_BACKEND[]:[]droom
[2010-01-25 21:11:54] 2010-01-25 21:11:54.851 MythSocket(ffffffff9862eae0:125): write -> 125 4       TRUE
[2010-01-25 21:11:54] 2010-01-25 21:11:54.851 MythSocket(ffffffff9862eae0:125): DownRef: 1
[2010-01-25 21:11:59] 2010-01-25 21:11:59.495 MythSocketThread(sock 0xffffffff9862eae0:125): socket is readable
[2010-01-25 21:11:59] 2010-01-25 21:11:59.495 MythSocketThread(sock 0xffffffff9862eae0:125): socket closed
[2010-01-25 21:11:59] 2010-01-25 21:11:59.495 MythSocket(ffffffff9862eae0:125): state change Connected -> Idle
[2010-01-25 21:11:59] 2010-01-25 21:11:59.496 MythSocket(ffffffff9862eae0:-1): calling m_cb->connectionClosed()
[2010-01-25 21:11:59] 2010-01-25 21:11:59.716 MythSocketThread(sock 0xffffffff9863fe68:127): socket is readable
[2010-01-25 21:11:59] 2010-01-25 21:11:59.716 MythSocketThread(sock 0xffffffff9863fe68:127): socket closed
[2010-01-25 21:11:59] 2010-01-25 21:11:59.716 MythSocket(ffffffff9863fe68:127): state change Connected -> Idle
[2010-01-25 21:11:59] 2010-01-25 21:11:59.717 MythSocket(ffffffff9863fe68:-1): calling m_cb->connectionClosed()
[2010-01-25 21:12:18] 2010-01-25 21:12:18.798 MythSocket(ffffffffa44cfb98:-1): DownRef: 0

When restarting a frontend on the 26th - I did notice an warning saying
something like :
frontend protocol is version 50 - backend protocol is (null/unknown)
Also on the other frontend I tried to get system status up - it didn't
have anything come up. 

MY SETUP INFO:

My FE/BE is latest 0.22fixes [23295] via svn on ubuntu intrepid. I have
1xdvb-T and 2xdvb-S adaptors with all EIT UK

Another FE is ubuntu karmic(ish) using packages from mythbuntu weekly
fixes. No hardware problems or other software problems AFAIK - load
isn't very high. No comm flagging and only a little bit of transcoding
recordings for my phone every now and again. 

Database seems to be fine and I've run optimise scripts recently (which
did speed up a few things). It is however the same db from around 0.16
or 0.18. Checked and it seems to be in latin charset still (mysql -e
status)

Thanks in advance.