Docker Hub tag mapping to Github Releases/Commits

[viharm]

Hi, I absolutely love Huginn and am grateful to the developers for this amazing tool. The rationale of naming it as "Huginn" is also fantastic!

I have deployed Huginn as a Docker container, and would like to understand the developers' view on the maturity of the image tags Docker Hub.

It appears as though there is an automation (huginnbuilder) in place to create a Docker Hub image from each commit on Github - the Docker images are tagged with the Github commit hashes. However the last "official" release on Github (as of 5-Mar-2023) was in Aug-2022 (tagged v2022.08.18).

To reduce risk to my production environment, I'm still using Docker image e605da6bd3e8cc49d94cc0e0f70ba488fb4f3fb2, which appears to be linked to this last official release.

So the question is - are the incremental "commit-based" Docker images considered to be stable? Or should we wait for images linked to "official" Github releases?

Thanks again for your hard work in developing such an amazing tool!

[viharm]

bump

[viharm]

Hi, I'm still interested to fully commit to using Huginn, I already have 93 agents, but it would be good to know this; as I'm sure there have been significant improvement since the last official release. Thanks.

[BlueHatbRit]

I've been running that last release using the dockerhub image probably since it came out. I came to update this week and had the same revelation you've had about the releases.

I decided to take a DB dump and "yolo" an upgrade. Upon doing so, I got a banner message in the app directing me to the GHCR image version which I've also switched to. Everything seems pretty stable for me at the moment, and I don't seem to be getting any errors from any of my agents. I can't speak for what your experience would be like though.

I don't have much else that I can add but here are my dump and restore scripts for a docker-compose deployment which might help you try an upgrade out and rollback if needed.

#!/bin/sh
# backup.sh - uses pg_dump to create a sql file with a dump of your database via
# the postgres container.

docker exec -i huginn_huginn_postgres_1 \
  /bin/bash -c "PGPASSWORD=DB_PASS pg_dump --username DB_USERNAME DB_NAME" \
  > ./db_dump.sql

#!/bin/sh
# restore.sh - uses psql to run a db_dump.sql file generated by pg_dump via
# the postgres container.

docker exec -i huginn_huginn_postgres_1 \
  /bin/bash -c "PGPASSWORD=DB_PASS psql --username DB_USERNAME DB_NAME" \
   < ./db_dump.sql

In both scripts you'll need to replace the following:

• huginn_huginn_postgres_1 with your postgres container name
• DB_PASS with your PG password
• DB_USERNAME with your PG username
• DB_NAME with your PG database

If you end up using the restore script, make sure it's on a fresh database, or amend the dump script as needed to account for an existing DB.

Hopefully you can use that to try out an upgrade and see how stable it is for you, and rollback if needed. It's been fine for me so far, but I've only been running it for a few days.

Assuming this works okay for me over the next few weeks, I'll probably just take an upgrade a month or something myself. I don't need uber stability or anything thankfully though.