thanks for your help, Joe. Yes, I have check the afr translator. In my opinion, "peer restart" operation does not trigger the process of keeping data identical, but the "read/write IO" in vm does trigger that process. if I am wrong, could you give me some advice to make sure the relation between "peer restart" operation and keeping data identical, how does the operation trigger the process afr translator did. thanks a lot, yz -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20160320/66856d9a/attachment.html>