This message was deleted.

OPA doesn't have any persistance, unless specifically configured via the experimental disk option . Not sure what you are experiencing. Can you share more info? Perhpas you are just seeing OPAL load the data quickly for you after the restart?

I just figured it out - it's pulling it from the policy repo! I happened to have a directory named

public

containing files called

data.json

and

huge_data.json

and it's loading the one named

huge_data.json

even though I have no mention of that file in any code.

| 15 | opal_client.policy.updater              | INFO  | Got policy bundle with 1 rego files, 1 data files, commit hash: 'f44737210315412e7fc9274e7a52c663afee2e99'

Yes :) https://docs.opal.ac/tutorials/track_a_git_repo#2-opal-client-subscribes-to-policy-update-notifications

Ooh - and the fact it's loading

huge_data.json

- is there a wildcard like

*data.json

or similar in place?

Yes 😇 Below the specific link are env var and manifest configuration instructions to control that, if you want

Perfect - for our actual deployment we'll have a repo just for the policy, it's just while testing I happened to have everything in one repo and was very confused when OPA was starting up with a ton of data in it as I'd missed the fact that you could load data from the policy repo not just rego policies. Mystery solved! Thanks for the help 😊

@Ben Wallis BTW - there is a new persistence feature in opal-client called “Offline Mode”. It loads OPA’s data from backup at start (when running OPA inline). That’s useful for not being dependent on server’s availability on client’s restarts. But it’s off by default. See docs for more details: https://docs.opal.ac/getting-started/running-opal/run-opal-client/opa-runner-parameters#policy-store-backup