Resource Stack: Choosing "PeopleSoft Cloud Manager for OCI" from the Marketplace.
In the end, this seemed to work reasonably well. Although I think I had some issues with this, and I ended up with three different VCNs with the same name.
Issues with installing the second IH (for self update) in a different VCN. See above.
- I ended up terminating the instances and doing the stack resource update again with a different VCN to avoid confusion.
- I should have really paid more attention.
The Linux server wouldn't create.
- After help from Oracle and trial and error (experimenting with creating instances manually) the issue seemed to be the that my account/tenancy didn't have the authority to create the instance using the Linux 7.7 template from the drop down in Cloud Manager.
- I had to find the 7.8 image from https://docs.cloud.oracle.com/en-us/iaas/images. This failed again due to selecting the Gen 2 image, so had to use the regular image.
The Windows provisioning failed (although the instance was created) as the 8.57 tools didn't exist on f:/somewhere.
- I had to subscribe to the Windows 8.57 Linux channel. Then redo the step in Cloud Manager.
The change package failed (Change Assistant), but you don't get much from the Terraform logs.
- I had to RDP to the server and look at the Change Package logs.
- An App Engine was trying to insert into a table where the unique key already existed.
- I marked the step as complete and resubmitted the job and it completed.
- I restarted the steps in Cloud Manager
To be honest I had to mark several steps as complete, and I don't really think the self-update even worked.
- I checked some of the update/bug ids from the download pages against the PS_MAINTENANCE_LOG and I couldn't see any of them.
Final thoughts on the self update
This was a very painful process. I think there is a lot of work that could be done to make this a more 'polished' process, and less prone to (user) errors. I found the terminology sometimes also confusing.
I find the Virtual Box based PI a lot easier to install! But this is probably due to more practice.
Installing FSCM Image #35
After subscribing to the channel, and creating an Environment Template I was able to create the environment. Once you click done (or start) the process starts.
This Cloud Manager process actually provisioned the Linux and Windows Server compute instances, but the DPK failed.
After looking at the logs on the server (the terraform logs were not terribly helpful), it turned out the Oracle database wasn't able to be created due to some missing prerequisites.
- Perhaps due to the different Linux build and 8.58 or 19c?!
Eventually I found MOS doc 2651114.1 which seemed to be a pretty close fit with the issue I had and recommended checking some other logs, which once I did, confirmed the same issue. The solution was to contact your IT department and installing the missing package.
- E-PUM: When Installing an 8.58 Based PUM the Following Error is Seen: Inventory.xml NOT Found (Doc ID 2651114.1)
Obviously I can install the package, and probably even start the puppet apply again. But I assume this won't feed back into Cloud Manager.
So ideally it needs to be done before the automated puppet build starts, i.e. as part of the DPK build. You can't do it earlier, as the creating the environment in Cloud Manager also provisions the Server.
Luckily, you can configure a script to be run before (and after I think) the DPK builds. This can be configured in the Environment Template or during the Environment build itself.
- I chose to do it in the template.
I then had to delete the environment again, which thankfully also terminates the instances in OCI as well.
After added the environment again, it (eventually) completed. This must have taken around 2 hours.
But in the end happy to have the new EP92U035 system up and running.
References
- For setting up Cloud Manager and provisioning an environment:
- For the self update:
