In brief:
When a postgres source connector is configured, it does not sync Postgres sequences. It will detect and sync tables, however this results in default auto-increment IDs not to populate, resulting in errors trying to write to the table.
TimescaleDB version: 17.7
PostgreSQL version: 2.24.0
Other software: N/A
OS: Debian
Platform: Tiger Cloud
Install method: Terraform
Environment: Development
We use TimescaleDB with Django. As part of the Django application, we use django-health-check to check the health of the DB, among other things. It does this by writing a test record to the DB before deleting it. On the source database, the testmodel table has the following structure:
Table "public.health_check_db_testmodel"
Column | Type | Collation | Nullable | Default
--------+------------------------+-----------+----------+-------------------------------------------------------
id | integer | | not null | nextval('health_check_db_testmodel_id_seq'::regclass)
title | character varying(128) | | not null |
Indexes:
"health_check_db_testmodel_pkey" PRIMARY KEY, btree (id)
Publications:
"925df574_4f58_45ca_b7cf_8a4b84682997_pub"
However, on the destination, with the postgres connector configured, the table structure is missing much:
Table "public.health_check_db_testmodel"
Column | Type | Collation | Nullable | Default
--------+------------------------+-----------+----------+---------
id | integer | | not null |
title | character varying(128) | | not null |
No indices or defaults. The sequence does not exist for the default to pull from:
Did not find any relation named "health_check_db_testmodel_id_seq".
The connector reports that it is running and there are Error tables (all in the _timescaledb_internal schema). There are no hypertables in use:
ERROR: relation "_timescaledb_internal._hyper_1_10_chunk" does not exist (SQLSTATE 42P01)
ERROR: all replication slots are in use (SQLSTATE 53400)