Closed Rub21 closed 7 months ago
@danrademacher - how does this fit into our current priorities?
We need more info from @Rub21 and @batpad here about the near term benefits. If the DB update will solve or mitigate issues like #642 and #633, then it should be a top priority, above OSMcha and coastlines.
If this is more of a long term tech debt fix, then we should stay focused on fixing 642 and 633, both of which directly affect mappers' ability to map.
Upstream has moved to PG 14, and PG 11 is end-of-life. So to match upstream, we would need to upgrade PG.
We would have to prepare for this and plan for potential downtime during the transition. Likely an hour or so. Or possibly we could do something more complex with a replica
Last time was something like:
Ruben has already been testing PG upgrade separately, and we could move forward with this.
the migration has been completed https://github.com/OpenHistoricalMap/ohm-deploy/pull/293
Per voice with Sanjay and Bitner, our database version 11 is a bit outdated and needs to be updated. According to the latest changes in the OSM website, seems the API software can be use with Postgres 14 github-repo: https://github.com/openstreetmap/openstreetmap-website/blob/master/docker/postgres/Dockerfile.
Also, updating the database will result in better performance, as it includes many improvements for performance, according to Bitner.
Since our last incident in DB performance, we have started monitoring the performance of the database since the number of users and the number of edits are increasing. It is necessary to improve the configuration and versions. and here is the results, that still need to be completed: https://github.com/OpenHistoricalMap/issues/wiki/OHM-Postgres-DB
cc. @jeffreyameyer @batpad @geohacker @danrademacher