[mythtv-users] Storage group mount is showing up twice

[Phil Bridges]

On Sat, Sep 6, 2008 at 7:56 AM, Mike Perkins
<mikep at randomtraveller.org.uk> wrote:
> Phil Bridges wrote:
>> Hopefully somebody can help with this.  I have a master backend (MBE)
>> and a slave backend (SBE).  I have two recording mounts in the MBE -
>> /var/video/recordings and /var/video2/recordings (which I just added
>> tonight) and one on the SBE - /var/video/recordings
>>
>> My MBE drives are locally linked at /mnt/MBE/recordings and
>> /mnt/MBE/recordings2.  My SBE drives are locally linked at
>> /mnt/SBE/recordings/
>>
>> On my MBE, the SBE drive is mounted at /mnt/SBE/recordings.  On my
>> SBE, my MBE drives are NFS mounted at /mnt/MBE/recordings and
>> /mnt/MBE/recordings2.
>>
>> My default storage group (the only storage group) ihas been entered on
>> the MBE and has the entries:  /mnt/MBE/recordings,
>> /mnt/MBE/recordings2, and /mnt/SBE/recordings.  There are no storage
>> groups defined on the SBE
>>
>> When I start up just the MBE, I look at the Mythweb status page, and it shows:
>>
>> MythTV Drive #1:
>>     * Directory: MBE:/mnt/SBE/recordings
>>
>> MythTV Drive #2:
>>     * Directory: MBE:/mnt/MBE/recordings
>>
>> MythTV Drive #3:
>>     * Directory: MBE:/mnt/MBE/recordings2
>>
>> Which is correct.
>>
>> When I start the SBE in addition, though, I get the following:
>>
>> MythTV Drive #1:
>>     * Directories: MBE:/mnt/SBE/recordings, SBE:/mnt/SBE/recordings
>>
>> MythTV Drive #2:
>>     * Directories: MBE:/mnt/MBE/recordings, SBE:/mnt/MBE/recordings
>>
>> MythTV Drive #3:
>>     * Directory: SBE:/mnt/MBE/recordings2
>>
>> MythTV Drive #3:
>>     * Directory: SBE:/mnt/MBE/recordings2
>>
>> Drive 3 is listed twice, both times with the SBE's mount point.  In
>> addition, the Total Disk Space entry lists the capacity of Drive #3
>> (the new drive) twice.
>>
>> Both backends have the following version numbers:
>> MythTV Version   : 18116
>> MythTV Branch    : branches/release-0-21-fixes
>> Library API      : 0.21.20080304-1
>> Network Protocol : 40
>>
>>
>> Does anyone here have any suggestions?
>>
> You have to be consistent across all servers with your naming convention, so
> that all partitions have the same names anywhere. The scheme I use (currently,
> for those who remember me sorting this out recently) is that every host has
> three mounts defined for it's own storage, and for every other (relevant) host,
> in the format:
>
> /mnt/<host1>/myth1
> /mnt/<host1>/myth2
> /mnt/<host1>/myth3
> ...
> /mnt/<host2>/myth1
> /mnt/<host2>/myth2
> /mnt/<host2>/myth3
> ...
>
> ...rinse and repeat, where <hostn> is the name of each of your mythtv hosts.
> These are where the partitions are mounted in /etc/fstab, and where the nfs
> share points are defined in /etc/fstab and /etc/exports. Note that this does not
> mean that one always has to have three partitions mounted, it just gives a bit
> of leeway for the future, and is reasonably standardised. (For those of you with
> more disks, set up more mount points. Obviously.)
>
> Under each real partition, create a directory 'video' which is what you add to
> your storage groups or whatever. This means that if for some reason the mount
> fails, myth won't attempt to write all over your root partition since the
> directory /mnt/<hostn>/mythn/video won't exist.
>
> (If you have spare mount points, it helps when debugging, since you can just
> swap the partition over to another mount point, and nothing breaks. E.g. myth1
> -> myth3, partition is still accessable to debugger, but myth won't attempt to
> write to it.)
>
> You can use variations of this method, but the rule is straightforward: each
> host must see every storage item with *exactly* the same name for mythtv to
> handle it correctly.
>

Thanks for the response.  They're all mounted like, this, so that
shouldn't be the problem.

Upon further inspection, my frontends list the drives correctly - it
looks like it's just Mythweb that is listing Drive #3 twice, with the
same remote mount point.