appropriate location for hammer snapshots

Siju George sgeorge.ml at gmail.com
Thu Feb 11 23:26:40 PST 2010


On Fri, Feb 12, 2010 at 12:00 PM, Matthew Dillon
<dillon at apollo.backplane.com> wrote:
> :Yes I checked it for both Master and Slave PFS the meta-data does get mirrored.
> :Some of them has 'migrated' at the end of it what does it mean?
> :
> :0x000000017627d370      2010-02-02 03:18:00 IST migrated
> :0x0000000176b5ef40      2010-02-03 03:18:00 IST migrated
> :0x00000001782a85e0      2010-02-04 03:18:21 IST
> :0x00000001782d0370      2010-02-05 03:18:17 IST
>
>    'migrated' is just in the comment field for the snapshot.
>    The comment field can be set when making manual snapshots
>    too.
>
> :is it possible to access the snapshot If I know the TID?
> :So as of now is it that I lost all the snapshots for my slave and I
> :have to manually  take snapshots of of the Slave PFS and store it some
> :where outside the PFS untill something is implemented for
> :regeneration?
> :...
> :Yes Matt thanks :-)
> :
> :--Siju
>
>    Yes, just use a similar path to the way the softlinks are
>    constructed, using the TID, e.g.:
>
>    cd <fs>/@@0x000000017627d370
>

Alright Matt. That is fine :-)

>    The snapshots are still there (pruning is governed by
>    the snapshot list in the meta-data).  If you want a
>    convenient set of softlinks you can construct them
>    by writing a little script based on the snapls output.
>

OK got it :-)

Thanks

--Siju






More information about the Users mailing list