when one container update fails, all other tasks are stopped/failed/unfinished

[underwoodblog]

As the title said, dockcheck.sh -p -u found multiple containers for update that where not related to each other. The last one failed (timeout of big slow download?) while downloading and all other failed too. worked when running each container update alone.

[mag37]

Thank you for raising the issue!

I agree, this is not pretty as you then manually need to do docker compose up -d on the containers that already had pulled, for them to use the new image.

A more reasonable thing would probably be to keep track of what pulls that failed and then not try to recreate those in the next step and finally end with printing what containers failed to update.

Maybe by changing up the docker pull section to something like..
First creating a new arrays SuccessfulUpdates=() and FailedUpdates=() earlier in the script.
Then:

      if docker pull "$ContImage"; then
        ### - redacted non relevant code - ###
        SuccessfulUpdates+=("$i")
      else
        printf "\n%bError pulling update for %S. Skipping. %b\n" "$c_red" "$i" "$c_reset"
        FailedUpdates+=("$i")
      fi

And then iterating the SuccessfulUpdates array to recreate the containers.

I'll look into it further as soon as I've got time!

[mag37]

( This could potentially tie in well with the #249 issue too. )