Termination of Ocim

Discussions - Public
21

@team Is there anyone knows anything about Login - OpenCraft - (charles.opencraft.hosting). The instance is still active and I’m curious if I can archive it.

22

@team I’m going to shut down the Ocim Production instance and the related infrastructure today. As no instances should be there which are used by any of our clients, this should not cause any issues.

I’ll do the termination of the infrastructure in three steps:

  1. Shut down Ocim Prod instance (but not terminate it) today and archive any other running instances
  2. Tomorrow, I’m going to shut down (but not terminate) the related infra (HAProxy, etc)
  3. On Wednesday, I’m going to terminate those resources that were stopped.

Why not wait until next Monday with termination? If I shut down the infra on Tuesday (i.e., tomorrow), it will turn out really fast if any of the instances are needed. So it is relatively safe to not wait more that 24h.

4 Likes
23

@gabor, since we no longer have stage HAProxy instances, should we remove these lines from our Ansible configs? I’m asking because every non-prod Open edX Ansible deployment fails when we try to reload the Consul (“Error reloading: Unexpected response code: 403 (Permission denied)”). The prod HAProxy instances seem to be online, though.

I didn’t open a PR, as I’m not sure if this won’t break the Node Exporter Ansible role. Do we still have a Prometheus stage instance that uses this data?

24

@mtyaka as you removed the staging env?

25

Yes, these should be removed, sorry for missing that.

I didn’t open a PR, as I’m not sure if this won’t break the Node Exporter Ansible role. Do we still have a Prometheus stage instance that uses this data?

No, we don’t have a stage prometheus instance as far as I know.

The prod HAProxy instances seem to be online, though.

Yes, the production HAProxy instances are used for a few things not related to Ocim, which is why we left them up. The stage instances were used exclusively by Ocim stage and we decided to remove them while decomissioning Ocim stage.

I’m asking because every non-prod Open edX Ansible deployment fails when we try to reload the Consul

I didn’t realize we were still using these playbooks to deploy Open edX. Which instances are we still deploying with ansible?