[Gluster-users] Gluster + QEMU + live Migration + cache != none

Jake Grimmett jog at mrc-lmb.cam.ac.uk
Thu Aug 9 11:27:23 UTC 2012

Dear All,

I've just updated my VM hosts to Scientific Linux 6.3. After setting the Disk 
cache mode to "default" (writethrough?) rather than "none" I get the following 
error when I try to migrate the VM:

migrating wiki1
error: Unsafe migration: Migration may lead to data corruption if disks use 
cache != none

Is gluster "coherent" across nodes, if so can I just use the --unsafe flag to 
force the move?

Many thanks,


More information about the Gluster-users mailing list