Questions re. Docker changes in 0.27.0

[tigattack]

I have some questions regarding the changes mentioned in the 0.27.0 release notes.

It's mentioned that the dawarich_db_data volume should be mapped to the app container for the Solid{Cache,Queue,Cable} SQLite DBs. Is the creation of SQLite DBs within the data directory of the PostGIS database intentional? It seems highly irregular to be working inside the data directory of another database for what appear to be purposes unrelated to PostGIS itself.

It's also mentioned that SolidQueue replaces Sidekiq and SolidCache replaces Redis. Does this mean the Redis container can be removed? And should the Sidekiq container be renamed? Or should the Sidekiq container even be removed, since it's noted that background jobs are run as a Puma plugin rather than in a separate container?

[Freika]

1. Yes, it's intentional. Postgres will ignore it anyway and due to the temporary nature of data stored there, even if it's somehow being purged (which it shouldn't) it will not become a big problem.
2. Both redis and sidekiq containers can be safely removed and it will be recommended in the nearest future. Background jobs are now being processed by SolidQueue instead if Sidekiq, and while it can also be run in form of a separate container, for self-hosting purposes it's not necessary and hence allows us to simplify the compose file

[tigattack]

Great, thanks for the answer! ❤️ I must admit I feel a little uneasy about the DB paths stuff. Even though PSQL will ignore it, I feel that sharing a volume in this way widens the scope for potential issues (and I really have the PSQL DB in mind here, more than the SQLite DBs holding temp data) when the cost of an additional volume is nil. All it would take is a small typo or other mistake in the entrypoint, for example, to cause serious damage to the PSQL DB.

[Freika]

I wanted to reuse what's already available 🤷 But you can create a separate volume and put your sqlite dbs there :)