Multiple errors when trying to run nhJoinCluster but the command seems to be successful.

Document ID : KB000047267
Last Modified Date : 14/02/2018
Show Technical Document Details

Problem: 

When trying to join a new system to an existing cluster I receive multiple errors about being unable to connect to ‘port 5050, host xxxxxxx’ or Host xxxxxxx: Error: Machine yyyyyyy is not a cluster member.

Environment:  

All current and supported eHealth versions and their supported operating systems.

Cause: 

The host listed in the above error was either offline (either full system or just the processes, use ps -ef | grep -i nhi to check for what is running) or something was blocking the bi-directional communication on port 5050.  If the nhJoinCluster command was in fact successful, the system that was having the problems would not have the new system on its cluster member table.

Resolution:

1. Resolve the communication problem (firewalls, ACL, etc) / bring the problem server online (nhServer start).

2. Go to the problem system and run the below command to get the cluster member table from existing and updated cluster members and update the cluster member table of the problem system to be in sync with the other cluster members.

nhUpdateClusterTable -cluster

Additional Information:

Example error when trying to run nhJoinClusterMember:


C:\Users\ehealth>nhJoinCluster -host F177316 -useStdDac yes -id 10005
Enter Existing Cluster Password: *******
Re-Enter Existing Cluster Password: *******
Join Cluster initiated:
 Validating supplied host .. OK
 Adding local machine to cluster .....Error: Unable to receive a message from
other process (unable to connect to 'port 5050, host F177317').
 OK
 Updating local database ................  OK
 Sending machine ID change notification.   OK
 Updating database handles .....Error: Unable to receive a message from another
process (unable to connect to 'port 5050, host F177317').
 OK
 Synchronizing objects of type group .... FAIL
 Synchronizing objects of type groupList .... FAIL
 Synchronizing objects of type groupListContents .... FAIL
 Synchronizing objects of type serviceProfile ..... FAIL
 Synchronizing objects of type dataAnalysisRec .. OK
Error: Unable to receive a message from another process (unable to connect to
port 5050, host F177317').
Host F177317: Error: Machine F177314 is not a cluster member.
Host F177317: Error: Machine F177314 is not a cluster member.
Host F177317: Error: Machine F177314 is not a cluster member.
 Replicating objects of type element .. OK
 Replicating objects of type group ..... FAIL
 Replicating objects of type groupSection .. OK
 Replicating objects of type groupList ..... FAIL
 Replicating objects of type groupListContents ..... FAIL
 Replicating objects of type serviceProfile ..... FAIL
 Replicating objects of type taReportUrl .. OK
Host F177317: Error: Machine F177314 is not a cluster member.
Host F177317: Error: Machine F177314 is not a cluster member.
Host F177317: Error: Machine F177314 is not a cluster member.
Host F177317: Error: Machine F177314 is not a cluster member.
 Enable schedule jobs for cluster admin .   OK

 

Pre fix: nhListClusterMembers -host F177317

Cluster List Members initiated:
 Querying cluster . OK
Cluster Members for host: F177317

MemberId    Name             MemberType   Trusted
+--------+--+-------------+--+---------+--+-----+
10004        F177310          server       Yes
Contacts:
10001        F177316          console     Yes
Contacts:
10003        F177317          server       Yes
Contacts:
10002        F177318          server       Yes
Contacts:

 

Post fix: nhListClusterMembers -host F177317

Cluster List Members initiated:
 Querying cluster . OK
Cluster Members for host: F177317

 

MemberId    Name             MemberType   Trusted
+--------+--+-------------+--+---------+--+-----+
10004        F177310  server       Yes
Contacts:
10005        F177314  server       Yes
Contacts:
10001        F177316  console     Yes
Contacts:
10003        F177317  server       Yes
Contacts:
10002        F177318  server       Yes
Contacts: