Add checklist for server migrations #955
Conversation
| - [ ] `provision.yml` | ||
| - [ ] `deploy.yml` | ||
| - [ ] `db_integrations` (Permit DB access for n8n, Metabase) | ||
| - [ ] Ensure sidekiq is disabled, to avoid creating subscription orders: `systemctl disable sidekiq` |
There was a problem hiding this comment.
Disable and stop in this case.
But I wonder if we should avoid copying production data before the switchover. 🤔
I guess it's part of the switchover and if problems are identified then we can fix it, go back and start again.
There was a problem hiding this comment.
Hmm I guess it's a question of if you want to separate out the testing from the switchover. Ideally, we wouldn't need to: everything just works and we setup, copy db and proceed to migrate and it's all good.
But we need to test and validate that things are working. And often, there will be something not working.
I'd rather identify that before co-ordinating and communicating a downtime, and putting the old site into maintenance mode.
Maybe for some smaller instances, we don't need that added layer of testing, and can accept more downtime while we frantically fix things quickly in-place. That might be a way more efficient use of time. (we ended up doing that for CA anyway!)
But I don't think that would be appropriate in all cases. Particularly for CA, I think it would have been too much to deal with all at once and we may have had to abort.
I have to admit I'm a bit scared of running db:reset, will need to double-check I'm definitely on the right server..
|
Thanks for the feedback @mkllnk , I've updated as per comments. |
|
I think this can be merged now, but I will gladly take further feedback. |
No description provided.