Symptoms
On attempt to navigate to OA > Infrastructure > Service Nodes the following error is displayed:
The operation you are trying to perform requires interaction with a component, which is presently unavailable. Please try again later.
The following error is displayed on attempt to view a databases page in CCP:
The operation you are trying to perform requires interaction with a component, which is presently unavailable. Please try again later.
Error ID: 1517397981970
No exceptions are registered in core.log. The below error is registered in /var/log/pa/pui/pui.log on the UI node:
11:26:21.948 [default task-92] INFO CORBA - #799544: org.jacorb.orb.Reference._is_a(user_id=213274 (su_user_id=1)) >>> entry
11:26:21.957 [default task-92] INFO CORBA - #799544: org.jacorb.orb.Reference._is_a(user_id=213274 (su_user_id=1)) <<< exit [9 ms] by exception: IDL:omg.org/CORBA/COMM_FAILURE:1.0
11:26:21.957 [default task-92] ERROR com.plesk.p2.wrapper.Handler - Rethrowing:
org.omg.CORBA.COMM_FAILURE: IOException: java.net.SocketException: Socket closed
at org.jacorb.orb.etf.ConnectionBase.to_COMM_FAILURE(ConnectionBase.java:152)
In /var/log/pa/console.log on the UI node:
180131 12:14:01 ERROR [jacorb.giop.conn] (default task-23) Underlying transport connection closed due to errors during sendMessage(), in ClientGIOPConnection to 192.168.10.2:8372 (62012f01)
180131 12:14:01 ERROR [stderr] (default task-23) org.omg.CORBA.COMM_FAILURE: IOException: java.net.SocketException: Socket closed
where 192.168.10.2 is an IP address of teh management node and 8372 is a port listened by the DatabaseManager service controller.
Cause
The DatabaseManager service controller was restarted:
Jan 30 23:34:51.820 : INF [0:4328:b4a63740 0:4328:b4a63740 pleskd ]: [ SCMonitor::doStartSC] doStartSC sc location_id: DatabaseManager:/usr/local/pem/libexec/DatabaseManager.so.7.0.1.11
but UI service were not restarted after that.
Resolution
Once service controllers on the management node are restarted, the UI services have to be restarted as well, in order to regenerate their security tokens. In order to fix the service controller availability issue, the UI services on all UI nodes have to be restarted:
# service pau restart