[mythtv-users] Slave BE startup failure

[Stephen P. Villano]

On 3/5/14, 10:09 AM, Raymond Wagner wrote:
> On Mar 5, 2014, at 8:15, "Stephen P. Villano" <stephen.p.villano at gmail.com> wrote:
>
>> My slave BE wasn't starting, it initially tests connectivity to the
>> primary BE, then complains that it can't connect to its own IP for the
>> database.
>> I'm on 0.27 Mythbuntu 12.04 with the latest patches.
> MythTV always tries to access the database first before connecting to the primary backend, except in the case where it doesn't know where the database is.  Then it will discover the master backend using UPnP, and query the backend for the location of the database.  If the master backend is configured to access the database from 127.0.0.1, it will tell clients to use that address too.
> _______________________________________________
>
Strange. In the my.cnf I used bind to the master BE IP (on the master
BE, of course). The debian.cnf uses the name localhost and in the
mythtv.cnf (in /etc/mysql/conf.d) it's 0.0.0.0.
So, something convinced mysql to not listen to the IP and apparently,
only listen to internal traffic.
Not too much of a biggie, I restarted the master BE's mysql and life is
good. If it happens again, I'll dash off a watchdog script to notify me
when the slave BE drops. It'll make tracing through logs a bit simpler.