7.2

7 Run the SQL query select * from databasechangeloglock.
The output should show a value of "f" for locked.
8 Start the horizon-workspace service using the command service horizon-workspace start.
Incorrect Appliance Role Assignments After Failover
After a failover occurs, master and replica vRealize Automation appliance nodes might not have the correct
role assignment, which aects all services that require database write access.
Problem
In a high availability cluster of vRealize Automation appliances, you shut down or make the master
database node inaccessible. You use the management console on another node to promote that node as the
new master, which restores vRealize Automation database write access.
Later, you bring the old master node back online, but the Database tab in its management console still lists
the node as the master node even though it is not. Aempts to use any node management console to clear
the problem by ocially promoting the old node back to master fail.
Solution
When failover occurs, follow these guidelines when conguring old versus new master nodes.
n
Before promoting another node to master, remove the previous master node from the load balancer
pool of vRealize Automation appliance nodes.
n
To have vRealize Automation bring an old master node back to the cluster, let the old machine come
online. Then, open the new master management console. Look for the old node listed as invalid under
the Database tab, and click its Reset buon.
After a successful reset, you may restore the old node to the load balancer pool of vRealize Automation
appliance nodes.
n
To manually bring an old master node back to the cluster, bring the machine online, and join it to the
cluster as if it were a new node. While joining, specify the newly promoted node as the primary node.
After successfully joining, you may restore the old node to the load balancer pool of vRealize
Automation appliance nodes.
n
Until you correctly reset or rejoin an old master node to the cluster, do not use its management console
for cluster management operations, even if the node came back online.
n
After you correctly reset or rejoin, you may promote an old node back to master.
Failures After Promotion of Replica and Master Nodes
A disk space issue, along with the promotion of replica and master vRealize Automation appliance database
nodes, might cause provisioning problems.
Problem
The master node runs out of disk space. You log in to its management interface Database page, and promote
a replica node with enough space to become the new master. Promotion appears to succeed when you
refresh the management interface page, even though an error message occurred.
Later, on the node that was the old master, you free up the disk space. After you promote the node back to
master, however, provisioning operations fail by being stuck IN_PROGRESS.
Cause
vRealize Automation cannot properly update the old master node conguration when the problem is not
enough space.
Installing vRealize Automation
136 VMware, Inc.