[mythtv-users] SMARTD reports a drive issue with a MythTV Storage drive

[Michael T. Dean]

On 01/17/2013 02:05 PM, Ben Kamen wrote:
> So SMARTD reported to me this morning that last night (probably 
> recording a show):
>
> CurrentPendingSector bumped up from 0 to 1
>
> and
>
> OfflineUncorrectableSector bumped up from 0 to 1.
>
> I run 2 drives for Myth but NOT in a RAID. So 1/2 of my TV recordings 
> are on one drive.
>
> What's the best method for backing this drive up so myth doesn't 
> complain?
>
> I would imagine it would be something like:
>
>   * Shutdown the masterbackend.
>   * copy all the files from the failing drive to someplace else
>   * replace with new drive and format/configure as needed
>   * copy files back to new drive from temp storage
>   * start up masterbackend
>
> Am I missing anything else?

That will work fine, though you don't need to  shut down the backend.

If you just want to prevent the backend from writing new recordings to 
the disk, do an "Archive" Storage Group and remove the directory on the 
failing drive's file system from any other Storage Groups (such as 
Default).  With this approach, you could still see the recordings on the 
failing drive (though this may be undesirable, as using the failing hard 
drive may speed its death?).  See 
http://www.gossamer-threads.com/lists/mythtv/users/535888#535888 for 
more information.

If you want to prevent the backend from writing to the disk or reading 
from the disk, just unmount the file system, then mount it at some other 
location that's not within any of your Storage Group directory lists.  
Then, MythTV will notice that the directory specified in the Storage 
Group directory list doesn't exist (may require a backend restart), so 
it won't use it.  Note that this assumes a proper configuration--where 
the Storage Group directory lists do not contain any "mount point" 
directories, but only subdirectories under the mount point.  (If the 
mount point directory is in the Storage Group directory list, it will 
still exist after the file system is unmounted, but it will be a 
directory on the parent file system--which will make a mess.)

Alternatively, you can just shut down the backend(s).

Mike