Hello everyone, following the (great) howto I've been able to get the MySQL cluster up and running flawlessy, but I found some problems configuring UltraMonkey. I followed the howto step by step (aka "cut and paste") and all check commands are replying as expected with a notable exception: if i disable ldirectord as a daemon I can't see weights shifting with "ipvsadm -L -n" (both servers are weight 0) like this: Code: IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.10.15:mysql wrr -> 192.168.10.11:mysql Local 0 0 0 -> 192.168.10.12:mysql Route 0 0 0 If i issue a "/etc/init.d/ldirectord start" weights start shifting as expected as one or the other server fails. My real problem is that if I stop the MySQL sql node on the active master I can't connect to the database anymore, thus not routing requests to the backup sql node. I can see in the output of ipvsadm the correct server weights (mysql stopped weight 0, the active one weight 1) Any suggestions? I can provide logs and/or conf files if you need them. Thank you for your time in advance
No, it doesn't report any error in logs... I found out that if I'm trying to connect from a third server the whole system works as expected. I take it that the service isn't balanced when one of the hosting servers is trying to access it. I have only 2 servers in my setup, with a web application (that resides on the same servers) accessing the database cluster. Do you think is possible to have a failover cluster with only 2 servers or I must have at least 3?
What IP address does your web application use to connect to MySQL? It should be the virtual IP address.
Yes, in my setup i have (192.168.0...) .11 and .12 as SQL interfaces and .15 as virtual IP for the balanced mysql to listen to. Now if i shutdown mysql on server .11 and try connect to server .15 i get the "unable to connect" error... Probably that's happening because of the lo:0 interface I configured