[mythtv-users] writeStringList error no data written on writeBlock

[Mark Perkins]

Hi List, hoping for some guidance on where to look to solve this one.

I have been experiencing something that I would term (probably incorrectly)
FE disconnects from BE (happens to both remote FE and BE/FE machine). It
seemed to only occur rarely, perhaps about once every 3 days or so (but more
often on the remote FE) and it seems to fix itself after about 5-15min
without any intervention. During this time Live TV does not work and
Recording Playback does not work.

However digging deeper there seems to be a lot more to this than something
occurring once every 3 days.

I am getting 1,000's (10,000's) of the following errors in the mythbackend
log file:

Aug 19 18:25:29 mark-P5K-Deluxe mythlogserver: mythbackend[500]: E
CoreContext mythsocket.cpp:375 (writeStringList) MythSocket(9d89210:159):
writeStringList: Error, No data written on writeBlock (622
errors)#012#011#011#011starts with: 122406
BACKEND_MESSAGE[]:[]GENERATED_PIXMAP[]:[]OK[]:[]1009...
Aug 19 18:25:30 mark-P5K-Deluxe mythlogserver: mythbackend[500]: E
CoreContext mythsocket.cpp:375 (writeStringList) MythSocket(9f10f80:157):
writeStringList: Error, No data written on writeBlock (603
errors)#012#011#011#011starts with: 122406
BACKEND_MESSAGE[]:[]GENERATED_PIXMAP[]:[]OK[]:[]1009...

They tend to come in blocks ranging from about 20-30min in length (at a rate
of 1 message per sec) and the block happens 3 - 6 times a day. It seems that
often when they occur we are not using the FE and so we don't notice
anything, recordings still appear to occur as per normal and scheduling /
housekeeping tasks all run as per normal. They just start in the log, at a
rate of about 1 per second, and after some 20 or 30 min or so they just
stop. I can't see anything that seems to occur beforehand that would make
them start or occur afterwards to make them stop.

However if we coincidentally try to do something with a FE (like watch a
recording or LiveTV) during one of these blocks this is when we see the
issues. FE will stall on recording loading screen for 30sec or so before
either dropping back to the program listing (usually), throwing the error
failed to jump to program buffer (or something similar) or freeze completely
(rarely). When it freezes completely it sits for a period of time before a
message pops up that the FE has crashed unexpectedly and has been
automatically restarted. There doesn't seem to be anything that can be done
to get LiveTV or recording playback to work from the FE other than wait.

Once whatever the problem is stops and the log file stops showing the
writeStringList error everything goes back to normal, FE will play recorded
programs and LiveTV. It's like nothing ever happened. No visible
interruption to recordings that are in progress during that period of time.

I just don't know where to start to track this one down. Any advice greatly
appreciated.

- Mythfrontend version: fixes/0.26 (v0.26.0-220-g92dbb43)
- MythTV Python Bindings
  local versions
    bindings version:         0.26.0.0
    ttvdb version:            1.2.1
    tmdb version:             v0.2.6
  external versions
    lxml version:             3.1.0
    MySQLdb version:          1.2.3.final.0
  protocol versions
    backend:                  75
    schema:                   1307
    music schema:             1018
    netvision schema:         1007

Mark