Search results

Correction : Seems you could relocate your CT with an ordered domain.

Hello, HA is meant to minimize downtime. Don't use HA ? As far as i know, there's no way to migrate CT/VMs automatically to one node or another. Maybe you could write a script if you really want it..

I remember why i ended installing glusterfs 3.5...I believe 3.4 didn't support "backup-volfile-servers" (with an "s") mount option. For now i reverted to 3.4.2 and it seems to work with libgfapi. root@proxmoxt1:~# qemu-img info...

With 3.5.0, glusterfsd processes crash randomly under load.. Well, 3.5.1 fixed a number of this bugs. I had no problems with reboots, but indeed use 3 nodes with replica 3. As seen on the official site "When client side quorum is enabled, there exists the rule such that atleast half of the...

Hello, I did not try glusterfs 3.4.2 with libgfapi. I wanted 3.5 for "Brick failure detection" and because i use one brick per disk (no raid), but it seems that release is not that good.. I guess there is a bug with qemu/libgfapi in 3.5 but i can't find it anywhere.. By the way, someone else...

Greetings, I have a working 3 nodes proxmox cluster. Using glusterfs as storage without problem (created a manual mount point + a shared directory.) Proxmox and glusterfs share the same nodes. Now i would like to use libgfapi and get rid of fuse overhead, but it can't get it to work.. I added...