Jasper Behrensdorf
Jasper Behrensdorf
I'm testing it on our cluster now to see if we can run 10000 samples in batches of 300 🤞.
I generalized the hpc functions a bit more. Any future hpc scheduler should override `setup_hpc_jobs` and `run_hpc_jobs` and that's it.
The extra deps are just for testing.
We're trying to put together a proper example using a model of a 42 story building in *OpenSees*.
I'm running `split-pro` in `docker` and push notifications work perfectly.
`split-pro` isn't using the `web-push` binary to send the notifications. It happens here: https://github.com/oss-apps/split-pro/blob/7f71b43c8f9fed8053ef76099542cde30b8ccb11/src/server/notification.ts#L12
Did you correctly set the environment variables: - `WEB_PUSH_EMAIL`, - `WEB_PUSH_PUBLIC_KEY`, - `WEB_PUSH_PRIVATE_KEY`?
No I see what your issue is. You're supposed to generated the `web-push` key somewhere else before running the container not inside.
I was thinking, maybe instead of verifying that the cluster connection works properly through the tests we export a function like this ```julia function test_slurm_connection(account::String, parition::String), ``` which will run...
I'm going to refactor the tests we have into a function that is passed a `SlurmInterface` and verifies that everything works as expected.