Workaround discovered, and is supported by VMware. Tested with vIDM 3.3.7 with latest patches and LCM 8.16.0.4 with latest pspak. This brings the cluster online within 15 minutes as opposed to anywhere from 45 to 90 minutes (or timing out when trying to start elasticsearch/opensearch) when powering on from LCM.
Run this on any node to make sure all 3 opensearch nodes are reporting: curl 'http://localhost:9200/_cat/nodes?v&s=cpu'
Find pgpool password in /usr/local/etc/pgpool.pwd – place in password location below
Run this on any node to make sure 3 postgres nodes are up and to find the postgres master: su root -c "echo -e 'password'|/opt/vmware/vpostgres/current/bin/psql -h localhost -p 9999 -U pgpool postgres -c \"show pool_nodes\""
Power off vm’s from LCM
Power on postgres primary node from vcenter
SSH to postgres primary and verify horizon-workspace service has started
Run (use netmask of eth0): ifconfig eth0:0 inet delegateIP netmask x.x.x.x
Run: /etc/init.d/horizon-workspace restart
Wait about 1 minute for this page to load on the primary https://host/SAAS/auth/login
Power on 2 other nodes in vcenter
Wait a couple of minutes for the SAAS login page to load on the other 2
Log into the vIDM VIP - Cluster health is green after a couple minutes
Run inventory sync in LCM
Enable cluster auto-recovery in LCM
