On Wed, Jan 06, 2016 at 08:53:31PM -0800, Matthew Ahrens wrote: > tricky code. In retrospect, we probably should have made the quota > enforcement even looser (more like user quotas), for the sake of > simplifying the code and making it less error-prone.
If it's not possible to prevent the live filesystem from going slightly over quota, and it turns out to be intractable to modify the recv code path to be as permissive about quota overruns, do you have any other thoughts on avoiding the failure situation we are running into where replication breaks and requires manual intervention to get going again? It looks like Dan's fix did resolve the issue for non-incremental sends, but I don't think sending over everything every time is going to be feasible :). Thanks... _______________________________________________ developer mailing list developer@open-zfs.org http://lists.open-zfs.org/mailman/listinfo/developer
