Edward Pilatowicz wrote:
> hey jerry,
> some final comments.
> ed
> - could you update the sn1 brand so that it will still work?
>   (it's broken on x86 because of 6703962, but it should still work on sparc.)

I'll take a look at what is going on there.

> usr/src/lib/libbrand/dtd/brand.dtd.1
> - so after reading the comments for "predetach" and "detach"
>   i still have no idea what the difference is between when
>   the two callbacks are invoked.

I had gotten similar input from Dan.  I have added more comments
to try to clarify this.

> - for the "clone" callback, is it possible for the clone operation
>   to fail after the callback has been invoked?  (in which case the
>   callback needs to be re-run friendly like the "*detach" callbacks.)

Yes it can fail, but it can't be re-run.  I think the comments already
explain this:

     If this hook exits with a non-zero exit status, the clone operation
     will fail and the zone will be left in the "incomplete" state, otherwise
     the state will be changed to "installed".

Let me know if you think I still need to clarify that.

> - "validatesnap", perhaps the description could be changed to:
>       Identifies the hook to invoke when cloning a zone to validate
>       the source zone snapshot that should have been created
>       using the built-in ZFS clone support.

I added some additional comments to clarify this.

Thanks again,

zones-discuss mailing list

Reply via email to