[OpenAFS] Feature request, sort of
Thu, 19 Jul 2007 17:10:07 -0600
Why not restore the volume (restores to RW), replicate it (same server
and partition) and then remove the RW?
Mount the resulting readonly explicitly -- i.e. be sure to include the
.readonly suffix in the fs mkm
Works for me.
Steve Simmons wrote:
> On Jul 18, 2007, at 10:27 AM, Carsten Schulz-Key wrote:
>> The RW and RO volumes share the same inodes (for the data part) as
>> long as
>> they're on the same partition and the RW volume has not been altered
>> releasing the volume. The space needed for the second volume header
>> should be
>> neglegtable. That's as far as I know -- please correct me if I'm wrong.
> You're correct; I'd not thought of putting them in the /vice partition.
>>> leaves two
>>> volumes rather than one that we have to remove later, and the savvy
>>> user could still mount the r/w original.
>> yeah, but you were going to delete it anyways, weren't you? So the
>> worst case
>> would be that the users mount the RW volume, delete the data and fill
>> it up
>> with other data -- which they can not rely on since it will be
>> deleted w/o
>> further notice a couple of days later.
> That's exactly the problem. Somebody sees they suddenly have twice as
> much space and starts using it - then gets quite upset when their new
> files all disappear. We'd like to have a mechanism that doesn't permit
> the user to go down that road. A read-only volume solves that, but we
> should be able to distinguish between a volume that is a standalone
> read-only and an abandoned replica. That, essentially, is my feature
> OpenAFS-info mailing list