[mythtv-users] my mythtv network stopped recording shows... why?

[stuart]

Hi...

Last week my mythtv network stopped recording shows.

I hit 100% on a partition (one of several located on both the MBE and 
SBE).  This despite a "don't record if you do not have 10GB available" 
rule.  I upped the rule to 30GB and manually (i.e. "rm <filename>") 
deleted some recordings.  I then ran the orphan program to update the 
data base (that is to remove the deleted programs from the data base as 
well).**

After 24 hours, mythtv still did not delete enough programs to satisfy 
the new 30GB rule on the partition that had been fill.  It was suggested 
that mythtv would only react to the new rule the next time mythtv tried 
to use the nearly full partition.  I then manually removed old programs 
my self using an instance of mythtvfrontend until the nearly full 
partition had at least 30GB of free space

(BTW:  I still would like an easy way to identify where a particular 
program is stored (i.e. computer:partition).  I have mentioned this 
feature request before - and people always recommended creating symbolic 
links using utilities outside of mythtv - then inspecting the links. 
This is fine in you are on a computer - but terribly inconvenient 
(reads: impossible) using only an instance of mythtfrontend.)

It has been a week - and it appears neither the MBE or SBE mythtv boxes 
have recorded a single show.  So I monitored the log files while 
manually instigating a recording while watching live TV.  I found this 
in the MBE log:

> 2009-06-28 10:15:48.192 Scheduler: AddRecording() recid: 265
> 2009-06-28 10:15:48.297 Reschedule requested for id 265.
> 2009-06-28 10:15:48.313 DB Error (UpdateMatches):
> Query was:
> DELETE FROM recordmatch WHERE recordid = 265
> Driver error was [2/144]:
> QMYSQL3: Unable to execute query
> Database error was:
> Table './mythconverg/recordmatch' is marked as crashed and last (automatic?) repair failed
> 
> 2009-06-28 10:15:48.382 DB Error (AddNewRecords recordmatch):
> Query was:
> UPDATE recordmatch  INNER JOIN sched_temp_record ON (recordmatch.recordid = sched_temp_record.recordid)  INNER JOIN program ON (recordmatch.chanid = program.chanid AND                         recordmatch.starttime = program.starttime AND                         recordmatch.manualid = program.manualid)  LEFT JOIN oldrecorded ON   (     sched_temp_record.dupmethod > 1 AND     oldrecorded.duplicate <> 0 AND     program.title = oldrecorded.title      AND      (       (program.programid <> '' AND program.generic = 0        AND program.programid = oldrecorded.programid)       OR       (oldrecorded.findid <> 0 AND         oldrecorded.findiPattern not found  (press RETURN)

...so, it looks like the data base is corrupt, people agree?

I can restore the data base to something that is 30 days old. 
Undoubtedly that will lead to orphaned files and more clean up work. 
Alternatively, are there any well known methods to clean up the data 
base with out moving back in time?  -  if in fact that is what is 
necessary here?

-thanks


** (Note: This 100% full partition on the MBE contained both a 
recordings directory as well as the MySQL data base.  I know - a very 
bad thing to do.  I promise the next incarnation of the MBE will not 
suffer from a similar set up!)