Who rated this post

Sharing in case it can be of use to anyone.

We installed a new ElasticXL/VSNext environment based on 9000 appliances, R82 T60.

First-time wizard, joining the cluster, creating VS/Interfaces/VSW, attaching them to VS, all good.

We then noticed frequent disconnections from the WebUI, switching between contexts in CLI is slow and so on.

CPView showed non-SND CPU were almost constantly at 100%, even with minimal traffic as production was gradually routed through the new VS's.

Further investigation showed that one of the nodes had CoreXL Dynamic Balancing disabled.

We enabled it in Local Clish, rebooted that member and after it was back in Active mode, CPU usage was back to normal ranges.