Did some playing around with LizardFS and used keepalived to setup a HA
solution for it - floating IP plus promote/demote scripts. The IP was
bound to a linux bond on two ethernet cards per node (balance-alb). The
proxmox cluster uses a thrid card per node with is a simple bridge onn a
sperate subnet.
It all worked well enough but I noticed after a while that the netwrok
started playing up - proxmox nodes dropped out of quorum, gluster
started having problems staying in sync with the heal count creeping up,
put I had no problem pinging any IP's and the syslog dind't show
anything odd. I disabled keepalived, rebooted everything and it came
right. This happened several times.
Are there any known compatibility problems with keepalived and proxmox?
I noticed all the other LizardFS has solutions used ucarp, but it seemed
less maintained and flexible than keepalived.
--
Lindsay Mathieson
_______________________________________________
pve-user mailing list
[email protected]
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user
