Tutor local reboot should build images before taking the platform down

I have a suggestion:

Currently, when I run tutor local reboot -d, the following happens (maybe more, but at least this happens):

  1. The currently running containers are stopped
  2. Images are pulled
  3. openedx and mfe Images are rebuilt
  4. Containers are started with the new/rebuilt images

Step 3. usually takes at least half an hour on my pretty decent VPS, so this means that the platform is completely unreachable and unusable for half an hour. That’s not optimal.

I think there wouldn’t be a problem in switching the steps around:

  1. Pull new images
  2. Build images
  3. Stop currently running images
  4. Start new images

This would take the platform down for only a few seconds, or a few minutes tops.

I usually try to run tutor images build openedx mfe before running the reboot command, but in this case I forgot to rebuild the mfe container and now I’m still “sitting with the baked pears”, as we say in Dutch…

When only images are updated, I suggest running only tutor local start -d. This should take care of pulling + building + running (in that order). If configuration files were modified as well, then you should run tutor local restart, which should be faster than reboot because it’s a docker-compose restart.