Joshua Moody
Joshua Moody
Not sure why harvester is not creating its own storage-class with the appropriate parameters for the harvester use case, then utilize that storage class for all usages? This way the...
On the longhorn side we should prevent the volume creation, if there is no node capable of hosting it during the api volume creation call. We could return the Out...
When I designed this, I had future extensibility in mind which you can see by the fact that there is a `CRYPTO_KEY_PROVIDER` field in the secret which allows us to...
@boopzz I think it's a valid ask to also have these mechanisms for any other type of Keys (Backups, ssl certs, etc). But since they require completely different implementations internally,...
@kompetenzlandkarte You can write a script around the below 2 commands, there might be some user scripts on github/slack for this already. If you do write a script around this,...
@kompetenzlandkarte you can also build something around the Python client, we use the same client for our e2e tests. https://longhorn.io/docs/1.1.2/references/longhorn-client-python/ https://github.com/longhorn/longhorn-tests/tree/master/manager/integration @yasker @innobead unfortunatly the lhexec script is setup to...
Some thoughts on this: - we should move the recurring schedule out of the volume crd into it's own crd - potentially just using the [kubernetes cronjob crd](https://kubernetes.io/docs/concepts/workloads/controllers/cron-jobs/) with labels...
Agree with @shuo-wu this requires some more thinking around the backup process. I like the idea of serializing some of the CRDs as part of the backup process since we...
@Echo0117 just allocate as much storage space for the volume as you want to maximal use for the filesystem. (You are effectively reserving this amount of space) Example you want...
The size of a snapshot (volume head counts as a snapshot) is bounded by the nominal volume size. So let's say your volume has a nominal capacity of 5GB, then...