If for some reason /var/pgdata.tmp already exists, probably because it wasn't cleaned from previous versions, Manager install fails on:
Install should either delete the existing file or use it but not fail on it.
Environment:
OS (CLI), HA cluster, cloud provider
------------------------------------
Steps to reproduce:
------------------
1. touch /var/pgdata.tmp on the leader of the cluster (simulate a failed failover from an earlier installation)
2. cfy cluster set-active manager2
3. cfy cluster nodes list
Expected result:
---------------
Waiting for manager2 to become the active node...
manager2 set as the new active node
All nodes are either in leader or replica state
Actual result:
-------------
Manager with /var/pgdata.tmp is in a failed state
Should be pretty easy to fix. Just make sure to delete the dir here: https://github.com/cloudify-cosmo/cloudify-premium/blob/master/cloudify_premium/ha/resources/sudo_trampoline.py#L126
Steps to reproduce:
1. touch /var/pgdata.tmp on the leader of the cluster (simulate a failed failover from an earlier installation)
2. cfy cluster set-active manager2