[mythtv-users] Slave backend disconnections after moving from r22901 to r22949
stichnot at gmail.com
Thu Dec 3 22:08:03 UTC 2009
I run trunk, and I recently moved from r22901 to r22949.
Subsequently, I couldn't get successful recordings from my slave
backend. After about 5 minutes (perhaps when commflagging reached a
key point), the master backend would think that the slave backend was
disconnected, with lots of messages like this in the log:
2009-12-03 06:04:52.370 MythSocket(ffffffffa99187c8:44):
writeStringList: Error, No data written on writeBlock (936 errors)
2009-12-03 06:04:52.371 MainServer, Warning: Unknown socket closing
2009-12-03 06:04:52.372 MainServer, Warning: Unknown socket closing
2009-12-03 06:04:52.374 MythSocket(ffffffffa9799018:-1):
writeStringList: Error, socket went unconnected.
We wrote 0 of 21 bytes with 1 errors
2009-12-03 06:04:53.372 MythSocket(ffffffffa99187c8:44):
writeStringList: Error, No data written on writeBlock (937 errors)
When this is going on, frontends have problems as well -- playback
stops responding to remote control commands, and the frontend appears
to hang while trying to load the list of recordings.
The slave backend has an HD-PVR which records onto NFS.
Today I reverted to r22901 and recordings are working again. (The DB
schemas are different by one version, 1246 & 1247, I can't wait to
deal with the fallout from that.)
I looked through the commits from r22901-r22949 but nothing jumps out
as obviously responsible for this problem.
I wonder if anyone else has seen this or has any ideas.
More information about the mythtv-users