Gluster users - Nov 2015 - Self healing slows down gluster clients file operations drastically

Hi all,

In our test lab we have a gluster distributed/replicated setup with 2 nodes
(1x2), a volume is created which is accessible by a glusterclient which lives on
a third node. We bring one node down in the cluster and write some files to the
existing volume. After that we bring up the node again and the selfheal daemon
starts healing all files nicely. Until here so far so good. But if we try to
delete all files again from the client while the self healing is still taking
place, the remove file command takes quite a long time to complete. If there?s
no healing process running the remove file command will execute in just a couple
of seconds.

It looks like the file that?s undergoing the healing process will get locked up
and is inaccessible to all it?s clients as long as the healing on that takes
place. Is this correct?

Thanks in advance.

Davy