Troubleshoot the Common service Health State status

A discrepancy may arise where, after you start the Common service, the Network Management Console (NMC) displays the Common service node as red, even though the oasErrLog.log file indicates that the service was started successfully. If you expand the Common service node, the child Health State is red. Although NMC is subscribed to a healthmon topic (<systemname>.common.healthmon.<machinename>healthstate), the NMC did not receive the update.

The above discrepancy occurs when, from the NMC on one system, you connect to another system and start the Common service of the standby machine immediately after the Common service was shut down. For example, as per the image below, an Administrator:

1. Starts NMC on the ls1pdc system,

2. Connects to the ls2pdc system, and

3. Shuts down and restarts the Common service on the LS2PDCRT2 machine.

   The child Health State is red.

   

When the Administrator starts another NMC instance on the ls1pdc system, the Common service node on the LS2PDCRT2 machine is green, indicating that the Common service started successfully.

In the above example, the PubSub Engine (PSEngine) on the Hot LS2PDCRT1 machine did not have enough time to finish the cleanup, reset the machine state, and declare the peer machine dead when the Common service on the LS2PDCRT2 machine was shut down. The issue would not have occurred had the Administrator started the Common service on the LS2PDCRT2 machine a minute or so later after shutting the former down.

The following configuration settings in two JSON files address this issue:

• Enforce a wait period if the Common service starts immediately after it was shut down:

  SystemServices.json

  {

  "SystemServices": {

  "PSEngineStartHoldOffTime": "00:00:40"

  }

  }

• Add a delay to ensure HealthMonitor publishes the <systemname>.common.healthmon.<machinename>healthstate topic after the machines recognize each other:

  ServiceFramework.json

  {

  "ServiceFramework": {

  "InitialHealthmonTopicPublishDelayMillisec ": 5000

  }

  }

AVEVA has tested these default settings in the R&D test environment and the issue was confirmed as fixed. However, depending on your system resources and environment, you may need to increase the values of these keys:

• Increase the PSEngineStartHoldOffTime key value (5 seconds increments) until the issue disappears.

• If you reach the 60 seconds mark and the issue is not resolved, increase the InitialHealthmonTopicPublishDelayMillisec key value (1 second increments) until the Common service node displays the correct state.