Our client has two “clustered” IS’s in Production and I have noted that once maybe twice a day the following messages are printed to the server.log:
[ISS.0033.0166I] Cluster Member Event: <????> has left the cluster.
[ISS.0033.0165I] Cluster Member Event: <????> has joined the cluster.
(the name of the IS has been replaced with question marks). Those two message codes are not listed in the Error Messages Reference guide.
Would anyone have a clue why this is occurring ?
Those messages appear when IS in cluster leaves or joins the the cluster. I guess IS name should be defined in extended settings to have meaningful name to appear in logs…
It might be due to network issue between the clustered nodes, or garbage collection takes more time. Do you have Java min and max memory value set to same?
Few more things that could probably cause this issue is:
- If the local address defined in the “Clustering” tab on the IS is not identified on the DNS.
- This most probably occurs when the physical server contains multiple network interfaces and one of them is nopt properly configured.
Yes I agree with Nanda. could be an issue with DNS not identified.
Thanks for the tips guys. The IS’s have been virtualised. I’ll check with the network guys to see what the DNS settings are… Oh, and I replaced the server name physically so as not to identify the customer.
I was told that this customer has done something funky and has disabled Garbage Collection in the JVM (?). Min is set to 1GB, Max is set to 6GB, but, that shouldn’t have any bearing on it should it ??
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.