Issue with many database services on RAC
Hi,
Has anyone experience with managing tens and hundreds of database services on RAC?
I have a 3 node cluster (Linux x86-64, 19.12) with ~200 services. Each service is active only on one node at the same time, other 2 nodes are set as "available" nodes, load is split between the nodes and only 1/3rd or around 70 services are active on each node. Failback is set to "yes".
The problem is that if any of the nodes is rebooted, one of the remaining nodes will be evicted. It happens always.
I reboot node a, services are being relocated to nodes b and c, database and services on node a is being stopped and after some time clusterware reports communication issues between remaining nodes and one of them is evicted.