Skip to main content

​
Before you begin

Take a backup of the database and API volumes before you begin the update process — the stack has to be stopped for the upgrade anyway, which is the same window a consistent snapshot needs. See Backup & restore. Check that the host has room for the new images. The pull downloads a complete new set before the old ones are released, so the upgrade needs headroom for both versions at once:

​
Steps

  1. Stop the current running containers
  1. Update the container images
If this step fails with no space left on device, the host has run out of room — usually because images from previous releases have accumulated. See Upgrade fails with “no space left on device” for how to reclaim it safely. The running deployment is unaffected, and the pull can be repeated.
  1. Re-create and start the containers