Proxmox GlusterFS

[tmatthaeus]

Hi everyone,


I have the following situation:


vz11 and vz12 are running proxmox and GlusterFS with a replica count 2 volume...


If I am migrate all VMs to vz12 and reboot vz11 then it seems, that not all changes which are made during the downtime of vz11 synchronized to vz11. So the VM Data between vz11 and vz12 is different...


I tried to trigger a selfheal with gluster volume heal vz11-vz12 full and with a find-operation like it is written on the GlusterFS website. But there seems no selfheal running.


What I am doing wrong? Which selfheal method do you use?


Or other question: What are you doing in a two node Proxmox / GlusterFS cluster, if you need to reboot one machine, or if one of this machine is broken temporarily?



And another question: How I can see, if the self heal process is still running or if the data are different?!


Thanks and Regards!

 

[vcp_ai]

Hi, I have a similar config as yours:

4 pve servers, 2 of them are running glusterfs server.​

Have a look at this post http://forum.proxmox.com/threads/18252-KVM-disk-corruption-on-glusterfs?p=93160#post93160
Where you can find two little utilities I use to see the status of glusterfs system.

Every time I do an operation like you describe (reboot one server), I try to assure that I have no machines that needs heal.

Also after restarting server, it takes a long time to get VMs disks in sync. So try to not have a high load on these machines during that time.

P.S. I do same as you describe everytime I update proxmox, and I've learned to be 'patient'.