Keepalived from catalog issue with rancher
volatilemolotov opened this issue · comments
Hello,
when i use this catalog entry i have an issue with keepalived. Failover works but when a crashed node is getting back up there is a brief period where it has no connectivity to the other node and presumes its the only MASTER and takes the VIRUALIP causing downtime. nopreempt and setting both to backup does not work because the two instances cannot communicate.
Any suggestions?