Open weaverba137 opened 2 months ago
It might also be a good idea to re-create the pg_dump
files. I recall there may have been minor changes applied to fuji
, guadalupe
and iron
after they were loaded.
As of today:
db2
.pg_dump
files for fuji
, guadalupe
and iron
have been created and backed up to HPSS.fuji
, guadalupe
and iron
have been loaded into db2
.And remaining to do:
db2
will be used to load kibo
and daily
.kibo
is loaded and tile-based loading has been tested for daily
, db2
will be renamed to db
, at which point it will take over as the official specprod-db
server.
The current production specprod database is based on PostgreSQL 11, which no longer supported. I have already created a Spin image based on PostgreSQL 16 and tested it on the Spin development cluster. The Dockerfile and related files are included in PR #14.
To avoid extended downtime on the existing production database, set up a separate database instance with persistent storage, load
fuji
,guadalupe
andiron
frompg_dump
files, then proceed to loaddaily
&kibo
.