[Bitkeeper-users] disk space and s.ChangeSet

[Mike Marty]

> > A user in my group ran out of disk space.  He did a checkin with the
> > citool and then did a pull.
> >
> > We aren't exactly sure when the disk space actually ran out, but the pull
> > complains that the s.ChangeSet is empty (which it is).
> >
> > We can retrieve a backup of s.ChangeSet from our nightly backup system.
> > But depending on when the checkin occurred, things might be out-of-sync.
> >
> > Any ideas?
>
> We can help you through this, it can be easier or hard depending on how out
> of sync the two repos are.
>

Thanks for the offer to help.  We decided it was easier to just restore
the nightly snapshot.

> A future version of bk will have a way of automatically fixing this when
> the fix is possible (i.e., the parent or some other repo has all the
> missing data).

I also might suggest that a future version have a more graceful way of
failing when disk space runs out.  We do use AFS and it was actually a
lack of disk quota, so that might have contributed.

(however using BitKeeper on shared AFS volumes has been nearly problem
free for the last 1.5 years, and we do _not_ use a server and instead
push/pull directly from shared directories)

--Mike