Hello,
I'm trying to understand the nuance of "repository" when it comes to restoring a snapshot on a DIFFERENT newly constructed stack. Here's the scenario:
- I create an ELK stack in US-EAST-1.
- I use the PUT/snapshot/my_repository with the { ... } approriate settings to create a respository
- I create a snapshot successfully in this repository
- I create an ELK stack in US-WEST-2.
- In this newly created stack in US-WEST-2 : I first try to get all the snapshots using:
GET /_snapshot/my_repository/_all
At this point I get an exception:
"type": "repository_missing_exception",
"reason": "[my_repository] missing"
However, if at step 5 BEFORE running that step in this new region, I do the step 2 (Create a repository with the SAME name my_repository using the same PUT API Call) and then attempt the GET all snapshots under the repository - things magically start working!
The Question is :
- Whenever we create a BRAND NEW ELK stack - is this create repository step mandatory ? If yes - is this like a "pre-registration" of sorts that is mandatory ?
- How would I account for this in curator if I create a new stack that starts up a curator job (as part of DR where I would need a new stack in a different region) ? would a manual intervention on my part be needed to run this "pre-registration" step needed before curator can do its job?