From what I've found out during my testing with rc2, the problem is how
replicate works. If you have two versions of files on remote1 and
remote2 the way they are synced is determined by your replicate1 config.
volume replicate1
type cluster/replicate
subvolumes remote1 remote2
end-volume
In this case, remote1 will sync to remote2, but not vise-versa.
Am I wrong group?
Ben Mok написа:
Hi All,
I am using 2.0rc1 with distributed + replicated, following by
http://www.gluster.org/docs/index.php/Mixing_DHT_and_AFR , all data
can be synchronized by “ls –R” after down server node is up. That’s
cool, no need to run self-heal script !! But I tested 2.0rc2 and rc4,
data can’t be sync and some files can’t be listed in client side after
all server nodes are up. I am using the same config files for testing
all versions, do I need to change something for rc2 and rc4?
Thank you so much !
Ben
------------------------------------------------------------------------
_______________________________________________
Gluster-devel mailing list
Gluster-devel@xxxxxxxxxx
http://lists.nongnu.org/mailman/listinfo/gluster-devel
