Skip to content

Commit b0389d1

Browse files
authored
bench infra: say that post-ingest.py retries through a database restart (#212)
1 parent 2297257 commit b0389d1

1 file changed

Lines changed: 3 additions & 2 deletions

File tree

‎infra/README.md‎

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -261,8 +261,9 @@ Automatic minor version upgrades are on. RDS applies a new Postgres 16.x
261261
release in the weekly maintenance window, Saturday 09:58–10:28 UTC. Single-AZ
262262
means the site and ingest are down while it restarts: about 30 seconds for the
263263
16.4 to 16.15 upgrade on 2026-10-28. An ingest job that connects during the
264-
restart fails, because `scripts/post-ingest.py` in `vortex-data/vortex` does
265-
not retry.
264+
restart waits it out: `scripts/post-ingest.py` in `vortex-data/vortex` retries a
265+
lost or refused connection for about 4 minutes, and each attempt writes the
266+
whole file again in one transaction. A restart that takes longer fails the job.
266267

267268
OS updates (`system-update`) are separate. A minor engine upgrade does not apply
268269
them, and RDS applies an optional one only when someone opts in. The OS update

0 commit comments

Comments
 (0)