[mythtv-users] MySQL related BE deadlocks - collective wisdom needed

warpme warpme at o2.pl
Fri Aug 26 08:39:23 UTC 2011

>> -in those 5 days period I had reported 1 DB successful reconnect.

>How are you accounting for that? i.e. what are you looking for in the
>log so that I can see how many I am experiencing.

Pls look for "MySQL reconnected successfully" string in be logs.
In last 12 days, 600 rec I had 2 of them. Both exactly at the same:

2011-08-15 06:00:01.626423 I  Tuning recording: "Jak to jest zrobione?":"Jak to jest zrobione? (How It's Made), odc. 2": channel 15315 on cardid 5, sourceid 8
2011-08-15 06:00:01.627545 I  MySQL reconnected successfully

2011-08-22 06:00:01.561354 I  Tuning recording: "Jak to jest zrobione?: Ołówki/Recykling metalu/Kawa": channel 15315 on cardid 5, sourceid 8
2011-08-22 06:00:01.562367 I  MySQL reconnected successfully

This is inline with symptoms of majority scheduler deadlocks I observed before redesign MySQL reconnect logic: I had bunch of "gueued" recordings and holding queue was opened exactly at 6:00AM.
It looks like there is clear correlation between 6:00AM, tunning to channel 15315 on cardid 5, sourceid 8 and DB reconnect. 
I think this might be sign of another very subtle bug.
Channel 15315 is very frequently used by me so thing is even more interesting that reconnect is always at the same time.
I checked sys logs & cron - no any relevant entries or jobs at 6:00AM. Interesting.


More information about the mythtv-users mailing list