[gpfsug-discuss] 220.127.116.11 changes
anacreo at gmail.com
Thu Jan 20 17:50:21 GMT 2022
Not sure if still supported but you could try tagging all the
files/directories in a fileset with the immutable attribute using the
policy engine. Not sure if that will cover what you need.
You can simply use policy engine to add or subtract that attribute at
will. It's biggest drawback may be exactly what you're looking for.
You could get fancy and mount the gpfs volume twice once read-wrire and a
second read-only.. then use mount command to mount over the read only
filesystem ontop of the read-write filesystem at the appropriate
junctures. You'd have to do this on each node you want affected...
Mount FileSys again as /ro/FileSys and share out /ro/FileSys/FileSetA use
read only option on mount.
Mount /ro/FileSys/FileSetA ontop of /FileSys/FileSetA.
I dont have root on my GPFS filesystems so I can't test this out... But
would work on most other Unit FSs.
On Thu, Jan 20, 2022, 3:10 AM Billich Heinrich Rainer (ID) <
heinrich.billich at id.ethz.ch> wrote:
> I just wonder if I can set a fileset to readonly - or link it as
> readonly. When I do data migration or finished a project which has it’s own
> fileset assigned it would be nice if I could just set the fileset to
> readonly to make sure that nothing and nobody can do any changes. There may
> be more cases …
> Immutability seems overkill, and I can’t revert it
> Related: If I setup an afm cache fileset in readonly mode – is this
> enforced on the cache side or just by convention? The afm gateways need to
> write to cache anyway, so probably other processes can, too?
> If anybody could comment and maybe also explain why it’s been designed the
> way it is?
> Thank you,
> gpfsug-discuss mailing list
> gpfsug-discuss at spectrumscale.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the gpfsug-discuss