If all else fails, go to your agent's data/ directory and look inside failover.dat -
remove the name of the bad server(s) in there.
Here's how ideally this could have been automatically fixed by RHQ:
1) You have multiple RHQ Servers in your RHQ HA Cloud, each agent has been given failover
lists with those servers (stored in failover.dat)
2) You change an RHQ Server IP, rendering one of the servers in the failover lists to be
invalid
3) The agent will fail to connect to that bad IP, switch over to another "good"
RHQ Server IP and continue running
4) After about an hour, all agents would ask their "good" servers for their
updated failover lists - this would remove the bad IP from their lists
I suspect you only have one RHQ Server in your environment, or you changed all of your RHQ
Servers and all the old IPs are bad.
In any event, <agent-install>/data/failover.dat can be manually manipulated in an
emergency (we don't mention this much in the documentation because that file is
supposed to be hidden from the view of users). I think you might have to restart the agent
if you touch that file manually.
----- Original Message -----
I think I have found something...
When I have installed the server, the IP address has a name in the hosts
file, ant this name hes been advertised to all of the agents. This name
doesn't exists any more.
How can I handle this? Many agents cannot connect at the moment. I have
repaired one by hand adding the old name again.
If I change the servers name, will the new name be advertised for the working
agents at least?
Attila
_______________________________________________
rhq-users mailing list
rhq-users(a)lists.fedorahosted.org
https://lists.fedorahosted.org/mailman/listinfo/rhq-users