mirror of
https://github.com/paperless-ngx/paperless-ngx.git
synced 2025-11-21 04:36:53 -06:00
c3ac102eba
This alters the retry/backoff logic in the init-wait-for-db script to be more optimistic about database availability. During regular deployment and operations of paperless-ngx, it's common to restart the application server with the database instance already running, so we should optimize for this case. Instead of unconditionally delaying 5 seconds between each connection attempt, start with a minimum delay of 1 second and increase the delay linearly with each attempt, maxing out at 10 seconds. This makes the retry count-based failure mode less practical, so instead we just use a timeout-based approach.* *NOTE: the original implementation would have an effective timeout of 25s. This alters the behavior to 60s. Additionally, this removes an unnecessary 5s delay that was injected in the postgres case. The script uses a more comprehensive connection check for postgres than it does mariadb, so if anything this 5s delay after getting an "ok" response from the DB was extra unnecessary in the postgres case.