search

reanahub / reana-workflow-controller

REANA Workflow Controller
http://reana-workflow-controller.readthedocs.io/
MIT License
2 stars 38 forks source link

db: REANA database gets deleted when cluster is restarted (PostgreSQL pod) #52

Closed anton-khodak closed 6 years ago

anton-khodak commented 6 years ago

Undesired behavior if a user wants to stop cluster or a development machine for a while

tiborsimko commented 6 years ago

Personally I am just suspending the VM when I'm not working on REANA, and resuming the same VM when I get back to working on REANA. Alternatively, you could also mount a local laptop folder into the minikube using minikube mount and use even hard delete/create cycle instead of the above suspend/resume cycle. Would these techniques address the original problem?

anton-khodak commented 6 years ago

By suspending you mean minikube stop && minikube start or...?

Yes, mounts would solve the problem, however, having a configuration to prevent analyses being erased after minikube stop, if feasible, would be more user-friendly.

Anyway, if the question of data persistence is solved in some other way in the production environment, this can be closed.

tiborsimko commented 6 years ago

I'm not suspending via minikube CLI, but via Hypervisor CLI tools, for example:

$ sudo virsh suspend minikube
$ # work on something else
$ sudo virsh resume minikube

The same can be achieved in VirtualBox GUI.

We should document this; moving the milestone.

diegodelemos commented 6 years ago

Title updated since this is now happening because the database disappears with the PostgreSQL pod. We should use persistent volume for /var/lib/postgresql/data.