As a fallout of T243609, we need to reset our Postgres DB to a sane state.
The process:
- -1) add option to puppet to disable tilerator
- 0) depool maps2004
- 1) isolate maps2004 from the rest of the cluster [MrG]
- 2) delete postgres data, recreate the empty DB [MrG]
- 3) run script to reimport data from OSM, needs to run from root [MrG]
osm-initial-import \ -d 200323 \ latest dump -p <password_file> \ (might not be needed) -s https://planet.openstreetmap.org/replication/hour/000/066/063.state.txt \ hourly replication -x webproxy.eqiad.wmnet:8080
- 4) disable tilerator
- 5) enable replication
- 6) check how replication is doing
- 7) repool maps2004
- 8) re-init all slaves (cookbook)
- 9) enable tilerator