Problem:
Following error is received when we try to join a new host system with Fullly Qualified Domain Name into an existing eHealth cluster:
nhJoinCluster -host <hostname with FQDN> -useStdDac yes
Validating supplied host ..Fatal Internal Error: Rcs timed out waiting for an ack for message 'CuIpcMsg'. Current state is 'needAck'..
(rcs/RcsSendProtocol::stateTimerCb)
Environment:
eHealth 6.2.2 and above installed on Windows / Linux / Unix operating systems
Cause:
The error means that provided hostname could not be contacted in time. It could either be due to network issue or DNS configuration.
Resolution:
As mentioned in the root cause, there are multiple factors that could be causing this error in cluster communication. To resolve this problem, follow the instructions mentioned below:
- Verify whether your cluster members are registered with either their fully qualified domain name (FQDN) or the short name.
- Confirm the hostname of the poller you are referencing in the -host parameter is correct, by using the following command on that machine
nhListClusterMembers
Make sure you are using the name you see in the list.
- Try to join cluster without using FQDN, if the above command returns name without domain name
- Do DNS lookup for the same hostname and ensure its reachable from this new system.
- Ensure DNS entries of all cluster systems are made in local hosts file