[mythtv-users] Auto repairing seektables
ismot at telemail.fi
Thu Jan 24 17:45:47 UTC 2008
On 01/23/2008 02:33 PM, Ismo Tanskanen wrote:
> I see quite often seektable errors in recordings. About one of ten
> recordings cannot be fast forvarded etc unless I run "mythcommflag
> --rebuild --file xxx.mpg".
> I'm thinking doing this automatically as cron job.
Better would be to fix the issue that's causing corruption. If it's not
a corrupt DB table, it's almost definitely an I/O wait issue. Is the
MySQL DB on a different filesystem (and, ideally, a different drive)
from your recordings drive? Is the root filesystem on a different
Note, also, that you may have much better luck with 0.21 when released.
> Is there any disadvantages if I make a script what will:
> - Drop mysql recorded-table
> - Run mythcommflag --rebuild to all files
Note that you would have to run mythcommflag --rebuild --file (or using
another approach to specify one show) on each and every file. Using
--all with --rebuild is not supported.
> Is recordedseek used for any other purposes? For example, will I loose
> commflagging points or something?
> Is it safe to drop this table and then
> do rebuild?
Semi-safe, but a waste (especially in a cron job). If you have 300
shows and each takes 5 minutes to rebuild... Rebuilding a perfectly
good seektable doesn't help.
Also, you're much better /never/ dropping tables. If you need to clean
one out, use "TRUNCATE TABLE recordedseek;" to keep the structure,
indices, ... and just get rid of the data.
mythtv-users mailing list
Thanks for answering, Mike.
How should I diagnose table/disk I/O problems?
I run optimize_mythdb regularly and it shows no problems. I have one
disk (ide 160gb) for system/mysql and another disk (sata 500gb) for
recordings. gkrellm shows very low disk utilization for system disk.
But, I have free unpartitioned space available on system disk. Should I
create partition for mysql database? Could it help if it is in same disk
I'm using svn 14912
More information about the mythtv-users