[Gluster-users] Strange behaviour in AFR - directory read only from first brick?

[Anand Avati]

>       As I test, AFR can migrate file (from Server2 to Server1 )which
> create at Server1 down
> time. But if a file is created when Server1 alive, then Server1 goes down
> (and all
> the data lose.) And restart Server1 (as well, glusterfs process), data can
> *not *be
> replicated by AFR automatically.


It is a little more detailed than that. If server1 disk is gone and you put
an empty disk in the backend, server1 is rebuilt properly. The issue is when
you take server1 offline, delete a few files  (and keep the directory
version intact) and plug it back in, the file can be rebuilt only when you
open the file by name (and will not be listed by ls). Handling this case
(only few files are deleted) is too expensive and makes normal operation
very slow. It is hard to detect changes without happening through AFR in
general (without compromising performance of 'normal' operations). We are
still on the lookout for a solution.

avati
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20080704/ccadda07/attachment.html>