Unable to upgrade from PAM 4.3 SP1 to 4.3 SP2
search cancel

Unable to upgrade from PAM 4.3 SP1 to 4.3 SP2

book

Article ID: 6747

calendar_today

Updated On:

Products

CA Process Automation Base

Issue/Introduction

Both domain orchestrators in the cluster were stopped

Services were stopped and set to manual, the servers even restarted, and the load balancer was showing the nodes down as well. 

On the installer screen for the load balancer configuration, the information for the node and the load balancer hostname were correct.

In the installation.log (located in /PAM/server/c2o/) the only information was the following:

<INFO  2017-05-17 14:32:54,956 DomainLoadBalancerScreen::info> Validating load balancer settings
<INFO  2017-05-17 14:32:54,956 OasisUtil::info> OasisUtil.checkHostName: nodeHostName: myserver.example.com

 

Environment

CA Process Automation 4.3 SP01 upgrading to 4.3 SP02

Windows Server 2012

Load Balancer: Any supported Load Balancer

Cause

The problem was that there were network issues on the servers (Proc Auto and Load Balancer).

Using the following URL, and substituting the load balancer hostname

https://[LOAD BALANCE HOST NAME]/itpam/ServerConfigurationRequestServlet?type=c2oServer&clusterHostName= [LOAD BALANCE HOST NAME]&process=clusterServer

This should return a page that states the website cannot be reached. 

What we saw instead was an error message from the load balancer that it was unable to resolve the address. 

This was due to the network issues in this environment.

Resolution

To get beyond this problem, in the hosts file on the primary Process Automation server, we added the IP address for a load balancer that was not having network issues with the hostname of the load balancer we are trying to connect with as:

192.0.2.1  myserver.example.com

So that the URL check will attempt the connection using 192.0.2.1 and get a positive response. In this case, the webpage is unavailable is a positive response vs unable to resolve.

From there, the upgrade continued, and the setting in the hosts file was removed after that check had passed. 

It's key to remember to remove the update in the hosts file before the primary node is started. 
No changes are required to move forward with the upgrade of the secondary node or nodes.