Ansible Automation Platform 2.5 Containerized Deployment

Hello all,

First time caller, long time listener. I am in progress of trying to modernize a deployment to AAP 2.5 on a brand new RHEL 9.4 AWS Instance. I followed the official documentation on how to implement it and cannot get past a certain point to make things work. Below is the hardware specs that are in line with the RHEL documentation from what I can tell, but if I missed anything then please let me know.

Redhat 9.4 latest updates as of 11/7
AWS EC2 T3.XLarge (4CPU, 16GB RAM, 100GB GP3 Volume 3000 IOPS)
Ansible Automation Platform 2.5 Containerized
SELinux Disabled

The main problem is that I got the installation to succeed following that, but I now have run into an issue to where no matter the job, inventory, or anything else I tweak, any template that I kick off immediately transitions into a “PENDING” state and will never move from there. I have changed almost anything I could think of and just cannot track down the why here. I attempted to follow the RHEL advice about how to fix that, but still no joy. Has anyone managed to get a setup like this working in AWS Land without any issues? Any advice would be much appreciated.

I had to make sure my postgresql db was version 15 or higher for the awx-operator

Thanks for that. I just checked my container installs, and I can see that I’m at version 15

[user@tower-containerized projects]$ podman ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
79df28c69a65 registry.redhat.io/rhel8/postgresql-15:latest run-postgresql 30 hours ago Up 7 hours
postgresql

I’ve gotten a very basic environment working now with a lot of the things I think working as designed. I’m now fighting a different issue relating to the PAH and collections, nothing seems to work as designed. Maybe I messed something up, or maybe it’s just broken, really not sure. Here is the link to the issue if anyone wants to follow along on this saga of riding the struggle bus.

http://forum.ansible.com/t/aap-2-5-collection-issues/10856

Pardon my ignorance, but if you purchase AAP, shouldn’t RedHat be responding to your problems via tickets or assigning some kind of installation tech to help or something?

I am in process with them as well on trying to configure things, but it never hurts to reach out to a wider audience and try to hopefully solve this issues that I’m facing with the wider wisdom of the crowd. I’ve just had bad luck somewhat with trying to get RHEL to respond in any kind of timely manner as it’s usually a days long wait for a response that is usually a question that kicks off another days long wait.

1 Like

Gotcha. Just academic curiosity about how worthwhile paying half a million dollars to RedHat pans out. Sorry, I don’t have experience setting up AAP, just AWX via kustomize.

AWX being what it is might still be helpful to figure out some things. Have you had any luck/issues trying to pull collections or anything like that from the PAH component, or do you not really do much with that. Main reason I ask is due to the fact that we have some custom modules that I have created that MUST be included in any EE as they are used in the bulk majority of our playbooks and I do not EVER see any official or community modules being released as this is such a niche software for the modules that even the vendor doesn’t make anything like it available.

The support may be more worth it if we paid for a higher level of service, but in this case it’s just the standard 8x5 support and so it takes forever there.

PAH stands for personal automation hub, right? I assume that’s their downstream version of Galaxy. Looking at your other thread, AAP 2.5 Collection Issues I noticed this

“https://localhost:8443/pulp_ansible/galaxy/published/. Got an unexpected error”,
“when getting available versions of collection test_collection.titan: Unknown”,

I have no idea how all this is meant to work with AAP since I don’t utilize it, but are you really meant to be using localhost when referencing your uh galaxy server? I only ask because later you referenced

that collection is configured with the following URL:

https://REMOVED:8443/pulp_ansible/galaxy/published/

You seem to know enough about technology enough for me to suspect you aren’t concerned about obfuscating localhost from us online strangers.

If you’re totally out of luck and/or just desperate to get things working, may be worth a shot to download the collections you care about from RedHat’s private registry and host them on your private/corporate Gitlab/Github server and reference them that way. Not a great long-term solution, but maybe something to get you going while waiting for RedHat to offer better help.

Sorry if this info isn’t helpful, I’ll stop replying so someone who has more experience can chime in.

yeah, I’ve done a bit with Ansible in the past and at least with the previous version prior to AAP it was no big deal to do most anything I wanted/needed to do. You are right that I wasn’t concerned about hiding localhost, the “REMOVED” was the hostname that was configured on the URL that was referenced and I didn’t want to leak any proprietary information if I could avoid it.

I would love to be able to download the collection from the RHEL registry, but that collection is that I developed myself and can’t upload it to the registry for much the same reason above as it’s proprietary company software as I created it while working here and just cannot make it publicly available. I will do some consideration about hosting it as a flat file on our github repo that all of our playbooks are stored in. I’m just glad that you’re attempting to throw two cents into things so that I can get some other eyes looking over it to make sure that I’m not being a complete idiot and missing something obvious. Don’t hold back on my account by no means.

Maybe upload your collection to your aap hub.

is your receptor container running?

Sorry for the delay. I’ve been tied up with some issues outside of this one. I FINALLY managed to do many things that I was having issues with. I was able to install everything as expected, build a custom EE, and even upload and install a collection into that newly built EE. sucks having to scour the internet to figure out the simple things that should have been added to the initial RHEL documentation.