Can I force an agent to come online?

Solved1.36K viewsFailover
6
0 Comments

I’m in a situation where both main and failover agents are offline. Can I force an agent to come online?

Marieke Goethals [SLC] [DevOps Catalyst] Selected answer as best 10th July 2023

1 Answer

11
2.98K 2 Comments

The SetDrsState SLNet Message is able to bring a DMA online (by force). This is to avoid disabling the Failover config but still be able to bring the DMA online. The option ForceEventWhenHeartBeatsFailing (bring a DMA online even when the heartbeats fail) combined with the BruteForce option. This option does the following:

  • Restarts the DMA that is stuck going offline (via the batch files).
  • Brings the DMA online without checking if the VIP is present.

This message should fix the following two scenarios:

  • When the DMA is in RTE trying to go offline and is stuck.
  • When another device has taken the VIP, we should still attempt to go online (however, this has some pitfalls; other DMAs won’t be able to communicate with this DMA anymore as the VIP has been duplicated): you will see notices that heartbeats and syncing are failing and the DMA won’t be accessible via the VIP.

You can find this message in the SLNetClientTest tool, by going to the “build message” tab page where you can search for this option, first I would do a ForceEventWhenHeartBeatsFailing; only if that one doesn’t allow the DMA to start up, you can use the BruteForce option.

Marieke Goethals [SLC] [DevOps Catalyst] Selected answer as best 10th July 2023
Jochen Dewachter [SLC] [DevOps Advocate] commented

Thanks! And can I do the opposite as well? Force an agent to go offline?
Marlies Vandenberghe [SLC] [DevOps Member] commented

If you really need an agent to go “offline” i would actually shut it down completely with the “DataMiner Stop DataMiner and SLNet.bat” file in the tools folder this will force the VIP to be released as well, if the other agent is brought online and the communication is OK it will automatically take the offline state when starting up again
If the communication is not OK the behavior will depend on the settings in the failover options. If you used inverted heartbeat it will go offline (cfr. chapter “Advanced Failover options” in the help).
However when doing this procedure on both agents when the connection is not OK, you will end up with both agents offline as well, the ForceEventWhenHeartBeatsFailing as described above allows you to chose which DMA will go online

As a side note, if the VIP can’t be released for some reason on one of the agents (this is normally the responsibility of Windows) you can only get this agent back online with the ForceEventWhenHeartBeatsFailing. The BruteForce will ignore the VIP and come online without VIP.