Mirrorless Cluster High availability services have issue

Products

• VMware Tanzu Greenplum
• VMware Tanzu Greenplum / Gemfire
• VMware Tanzu Data Suite
• VMware Tanzu Data Suite

Issue/Introduction

After one of the VM host reboot(graceful not a crash), the HA services did not automatically bring the database back online on that segment host on that VM host.

 

 

 

Environment

Mirrorless Greenplum Cluster

Cause

On graceful vm reboot the postmaster.pid file of the vm's gets cleaned which is used by the HA service to bring back the segment to its original state , how it started before.

Resolution

Work Around -

Restart the db to get the Cluster back to its original state.

 

Fix -

This situation is expected as only in case of VM crash / kernel panic, the postmaster.pid file will still exist on the filesystem, upon reboot postmaster service will restart that segment based on that pid

Additional Information

Do you need to remove the postmaster.pid file before manual restart of the segment ? 

If its a vm reboot, the reboot is graceful  the postmaster.pid file gets automatically removed when it done. So no its not required to remove.
For the second question