FPP by Example Part 2: FPP Server Installation

In the previous part of this series, we touched upon Fleet Patching & Provisioning (FPP) concepts and architecture;  it is now time to put the theory into practice.

As you recall from the previous post, the FPP architecture looks as described in the picture below:

With the review of the architecture fresh in mind we can start configuring the FPP Server on our already installed Grid Infrastructure. Any Real Application Clusters (RAC) installation ( single RAC node, multiple RAC nodes or RAC one node) and ACFS compatible kernel will do.

I will now walk you through the following steps to install the FPP Server, which will provide the foundation of our FPP configuration:

1. Install the MGMTDB if not already there
2. Remove the local rhpserver
3. Add gns ip
4. Add the rhpserver
5. Start the rhpserver

Step 1 – Install MGMTDB

FPP stores its metadata in the MGMTDB. Make sure that you have the MGMTDB installed.
If the MGMTDB is not installed you can use the tool our colleague Ruggero Citton from the RAC Pack wrote to add it to your configuration quickly. Please note this works for 19c, in 21c the MGMTDB needs to be in its own database home and can thus no longer use the GRID_HOME.

For more info check out the following MOS note :
MDBUtil : GI Management Repository configuration tool (Doc ID 2065175.1). Make sure ssh equivalence for the grid user is set up.

It is interesting to know that all 21c features of FPP are backported to 19c as this is the long-term release version. At the time of this blog post 19c would be the recommended version to run in most cases.

Step 2 – Remove the Local rhpserver

By default a local automation fpp server is installed, This local rhpserver need to be removed to install the full fppserver
 

As the grid user, run the following command:

Step 3 - Add GNS (optional as of 19.25)

Update : GNS is deprecated in 23ai. Therefore, we recommend using the RHP VIP as of version 19.25. Adding GNS can be skipped if you make use of the RHP VIP.

FPP uses basic GNS to let the FPP clients resolve its internal ip addresses, it is not a full blown GNS installation no subdomain is necessary only one IP address

As the root user, run following command:

Step 4 – Add the rhpserver

We removed the rhpserver earlier and need to re-add it, to complete the FPP server installation.

As the root user, run the following command:

This command will add the rhpserver and will create an ACFS filesystem in the +DATA diskgroup and mount it in /rhp_storage on the FPP server. The rhpsvip parameter is used when you don’t have GNS configured.

Gold images will be stored in:

As you know FPP is idempotent. Whenever an errror occurs, FPP operations will resume after correcting the issue assuming the same command is re-run.

Progress of FPP actions are trackin in following locations:

Step 5 – Start the rhpserver

With the rhpserver added we can now start it by running following command as the grid user:

With all the components installed and succesfully started it is time for a short sanity check.

By running the following command, you should see the following cluster resources

In the case above I have a two-node RAC Cluster with fpp19c-c11 and fpp19c-c12 as RAC Nodes.

Congratulations, if you have followed the steps above in you environment, you should now have your FPP server up and running !!

In the next blog post in this series, I will walk you through how to create and import  gold images.

You can find the links to other articles in the series here :

Fleet Patching & Provisioning by Example Intro

Fleet Patching & Provisioning by Example Part 1 : Architecture – Concepts

Fleet Patching & Provisioning by Example Part 2 : FPP Server Installation

Fleet Patching & Provisioning by Example Part 3 : Creating Gold Images

www.oracle.com/goto/FPP