[mythtv] Schema updates

Michael T. Dean mtdean at thirdcontact.com
Mon Mar 5 16:08:52 UTC 2007


On 03/05/2007 02:12 AM, f-myth-users at media.mit.edu wrote:
>     > Date: Mon, 05 Mar 2007 00:31:25 -0500
>     > From: "Michael T. Dean" <mtdean at thirdcontact.com>
>
>     > On 03/05/2007 12:20 AM, f-myth-users at media.mit.edu wrote:
>     > > [Do we repair/optimize the DB before allowing an upgrade to commence?
>     > > Should we?  Seems like a good idea---it might lessen the risk that a
>     > > crashed table will stop the upgrade in an unfortunate place, with
>     > > indeterminate results.]
>
>     > http://www.gossamer-threads.com/lists/mythtv/users/254631#254631
>
> Oh right.
>
> I remembered that thread, but was remembering it as "repairing and
> optimizing daily at random times" and forgot about R/O failing on
> non-MyISAM etc.  But maybe we -could- just ignore failures on other
> engines?  If we tried to do it automatically (I'm talking for schema
> upgrades only, here :), it'd be one less step for "normal" users to
> remember, and presumably people who've changed their storage engines
> from the default are on their own anyway.
>   

My only remaining concern would be the possibility of the MySQL server's 
crashing during a REPAIR--at which point another REPAIR should be run 
before the table on which the REPAIR was executing when the DB crashed 
should be run immediately (before accessing the data).  I still think 
that hiding a REPAIR is a dangerous thing (as someone who doesn't know 
it's happening has no chance of knowing what to do if something goes 
wrong), and Myth can't guarantee a REPAIR early enough to fix it 
automatically.

> (I admit it's not a crucial point, but ISTR a couple of people in the
> last few months having upgrades go awry because they started with
> broken tables---and then having lots of trouble debugging the
> resulting incompletely-upgraded state.  Sure, they shoulda fixed 'em
> manually first, but...)

Yeah.  It does happen.  However, if we could convince you and James 
Chapman to work on the embedded MySQL thing, the whole issue could 
disappear and the result is a much cleaner solution.  ;)

Anyone have any experience with using embedded MySQL with QT3 (as 
opposed to QT4)?

Mike



More information about the mythtv-dev mailing list