On Tue, 12 Apr 2005, Bruce Markey wrote: > > I too have seen this, I took a cvs snap from last night, just before > > bjm's last UNKNOWN_COMMAND, but there were others just before > > it that looked promising. I've only run 1 day on it so I can't give a full > > report but at least one time these messages seemed to precede a > > necessary restart of both master / slave backends. > > As I'd mentioned in another thread, the changes I've made are more > geared to trying to recover after something has gone wrong. We still > don't know for a fact what exactly triggered the out of sequence or > garbled messages though some of the things Isaac did probably address > the most likely causes. BTW, the Overlapping message is more debugging > info than any sort of serious problem. > > If you could, please do the following for the next day or so. Install > current CVS and run: > > mythbackend -v network >> /tmp/mb.log 2>&1 & > > on both the master and any slaves. Later "grep Protocol /tmp/mb.log". > If there is a match, please cut'n'paste all the messages from > a minute before up to a minute after the "Protocol error:" message > and grab the messages in the same time frame from the slave's log. > > Thank you, > > -- bjm
ok, I will build it up tonight. Thanks, Brian
_______________________________________________ mythtv-dev mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
