I am trying to doc this procedure but have not found the answer to this question yet.
The use case is a manual recovery from a failed Orion main server. All of the nodes are assigned to additional pollers (none are polled from the main Orion).
I have seen these step recommended but I think there is a missing step to point the pollers to the secondary Orion.
Any help is appreciated.
The Primary Polling Engine has to be online when you connect the Additional Polling Engine (unless you are on NPM v9.5.1 or less).
The Additional Polling Engines will authenticate with the Primary Polling Engine, then connect directly to the database server.
From here on after, the Additional Polling engines will only connect to the Primary Polling Engine if or when you need to assign a node to said Additional Polling Engine. Until then, the Additional Polling Engine will only connect to the DB.
1. Make sure the Primary Polling Engine is online and running.
2. You may need to run Conf Wiz if no connection between Primary and Additional Polling Engine is present. Check the Polling Engines status in Settings > Polling Engines in the web console.
3. As above, check Polling Engine status in Settings page. The run test like Discover, Poll Now, List Resources for nodes assigned to Additional Polling Engine(s)
I'm a little confused with this answer. The use case is that the primary poller has failed. How do we make the additional pollers link with the secondary? This environment is not using FoE.
PPE: Primary Polling Engine
APE: Additional Polling Engine
Additional Polling Engines connect to Primary Polling Engines via common database.
When you say secondary, do you mean a second Primary polling engine (main Orion NPM)?
If so, you will need to connect your Additional Polling Engine to the new/other Primary Polling Engine database server.
1. Make sure the second primary polling engine is online and polling
2. Run configuration wizard on the additional polling engine that you wish to connect to the second primary polling engine.
3. Perform poll assignments for nodes on the second primary polling engine in Manage Nodes > More Actions > Change Polling Engine
4. Perform tests.
Ok - that is much clearer. It seems step 2 above is the step my procedure is missing. All of my VMs are tied up or else I would test the procedure myself. The reason I question if config wiz would do the trick is that CW is typically only aware of db, web, and services. When you install the additional polling wizard the assignment to the main poller is done in the installer before CW is run. It seems odd to have the assignment made one way when you install and a different way after installation. I think the additional polling engine has to know the main poller IP to communicate with the main poller on 17777.
Is the reassignment of additional poller documented? I have been looking for a KB or something and can't find one. I'll reach out to Pacetti-san and see what he says.
Thanks for your help on this one.
"When you install the additional polling wizard the assignment to the main poller is done in the installer before CW is run"
Although this seems logically sound, this is not the way it is done. The polling engine checker or (where you type in the IP/Hostname of the primary server and login credentials) is only there to confirm version support. Connection to the PPE's Information Service is done through the Configuration Wizard. Things to note: The APE MUST be the same version as the PPE. If you know that these two are running with the same version software like for example 10.4.2 then you can safely go ahead and assign the APE to the new/secondary PPE.
APE's are only aware of PPE's through the database. That is where polling engine(s) details is stored.
Remember, after completing Configuration Wizard, you must re-assign your nodes to the correct polling engine. Documentation about poller assignments can be found in the admin guide on page 141-142 (latest guide as of today).
I think missing documentation is due to the fact that this practice not being the preferred method of solving a poller related issue.
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 150,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process. Learn more today by joining now.