IT Process Automation

Hey guys,

Is there a way to get the PAM workflows to keep running against a new host when you replicate the database?

Following this tecdoc : How do I configure a new PAM Orchestrator instance to use existing PAM database tables? 

It says : This process is not supported for in-flight processes.  Running processes contain links to the current Orchestrator and therefore may not continue to process after migration to a new host.  All running processes should be aborted before following these steps.

How do we work around that?

The only idea i'm having right now is :

Add a new node in my pam cluster on server A. Shutdown every other orchestrator to get every processes to run against my new orchestrator.

Backup PamRuntime. Remove node from cluster on server A. Restore backed up DB to new DB server. Reinstall PAM on server A. Link it to DB on new DB server. Follow tecdoc to change orchestraot id in config files.

Makes sense?

Sorry Pier,

This has been tried and will occasionally work, but that warning message was added after repeated failures to get instances to work properly after the Orchestrator ID changes.

In 4.3 sp2 we have enabled a 'disaster recovery' option, which allows for something like what you have described, except your new environment will be initially deployed as an inactive, passive environment until the main active environment is turned off.

You can get details in the documentation starting here:

How to Use the Disaster Recovery Plan

Okay this can solve my issue when migrating my prod ENV to new servers.

I understand that i cannot use this approach to replicate my prod ENV with active processes as i will have to shutdown my prod ENV.

Can i remove a node from a cluster without uninstalling it?

So i read the documentation and I'm having more questions

Cluster A is composed of Node 1, Node 2 and Node 3. (All on windows server 2008)

I create Node 4, Node 5 and Node 6 (All on windows server 2016). I assume i can update the datastore settings during the installation to point to the mirrorred database.

I shutdown Node 1, Node 2 and Node 3.

Node 5, Node 6 and Node 7 should start and become available for request. They should also keep the active processes running.

Now i want to remove Node 1, Node 2 and Node 3 from the equation because i want to end up with two distinct cluster setup, lab and prod. And when we will be migrating our prod env, Node 1, Node 2 and Node 3 servers will be deleted.

Any hints?

Hi Pier, wow that's  a complicated scenario!

So to get this thing working, you would have to select the Disaster recover option during installation/configuration in all nodes. Old and new ones. So when you do this, isn't the servers down?I mean, during the installation. 

If so, it answers to your question, no it's not possible.

You will have to set up a disaster recovery scenario and when you do so the servers are going to be down. At that point, you could just get rid of the nodes that are not going to be used, (node 1,2,3).

Thnak you Marcel