[mythtv-users] Broken recordedseek table - what next?

[Robert Johnston]

On 21/06/2010 2:28 AM, Nick Rout wrote:
> On Mon, Jun 21, 2010 at 6:55 PM, Robert Johnston<anaerin at gmail.com>  wrote:
>> On 21/06/2010 12:21 AM, David Whyte wrote:
>>>
>>> On Mon, Jun 21, 2010 at 4:06 PM, Nick Rout<nick.rout at gmail.com>    wrote:
>>>
>>>> What now? Is there a way to repair it?
>>>>
>>>
>>> I often run the 'repair' options from the MythWeb>    Settings pages
>>> when I get corrupt tables.
>>
>> If repairing simply won't work (Which it seems, from the log posted, has
>> been tried and failed), you could always drop and re-make the table, then
>> use mythcommflag to rebuild the seek tables. That would take a while, but
>> would ensure you have few problems.
>>
>> Of course, you could always use the backup that the cron job produced to
>> re-create the table, and mythcommflag to rebuild the missing seek table
>> entries. This would reduce the time spent rebuilding, but would rely on you
>> being able to pick the correct create/insert statement block(s) from the
>> backup.
>
>
> I have a backup that is a week old (today's failed, presumably because
> of the corruption).
>
> When I try either run the cron script, or do REBUILD TABLE
> recordedseek; within mysql's command line I get this stuff repeated in
> dmesg:

<snip type="Operating system crying out in agony" />

> This seems to resolve in google to hard drive errors, so I booted to a
> ubuntu desktop cd and ran fsck /dev/sda2 but it declares that all is
> well.
>
> sda is an older 20G IDE hard drive that i used to attempt to put the
> OS and database on a separate spindle. Maybe it is dying and the
> stress of the database rebuild causes a developing error to manifest
> itself?

Yes. What it's saying is that there is (at least one) bad sector on 
/dev/sda, and the drive has no more "reserved" sectors to swap it into. 
My advice would be to replace that hard drive with something, anything 
else just as fast as you can - it's about to go nuclear.