pg_last_xlog_receive_location() does not exist

[once-ler]

Version 7.1.0 breaks with

ERROR: function pg_last_xlog_receive_location() does not exist at character 283

Could this be the cause? PG 10 renamed "xlog" to "wal" as stated here

[sumedhpathak]

Hi there! Could you add some context on the situation when you run into this issue? Since this is in the Docker repo, I presume you encounter this in a Docker image. I don't think we're doing anything specific w.r.t the transaction log in Docker, so it may be a more general issue

[once-ler]

I am running docker-compose -p citus up

I then get errors such as worker_1 | 2017-12-13 20:25:40.868 UTC [1] DETAIL: The data directory was initialized by PostgreSQL version 9.6, which is not compatible with this version 10.1.

Previously, I had installed apt-get install postgresql-server-dev-9.6 on my desktop, but I don't think this would be relevant to docker.

Next, I tried docker version 7.0.1, since the base image uses PostgreSQL 9.6. The master and manager starts but the worker gets killed.

worker_2   | /docker-entrypoint.sh: running /docker-entrypoint-initdb.d/001-create-citus-extension.sql
worker_2   | LOG:  starting maintenance daemon on database 12407 user 10
worker_2   | CONTEXT:  Citus maintenance daemon for database 12407 user 10
worker_2   | CREATE EXTENSION
worker_2   | 
worker_2   | 
worker_2   | LOG:  received fast shutdown request
worker_2   | waiting for server to shut down...LOG:  aborting any active transactions
worker_2   | .FATAL:  terminating connection due to administrator command
worker_2   | CONTEXT:  Citus maintenance daemon for database 12407 user 10
worker_2   | LOG:  autovacuum launcher shutting down
worker_2   | FATAL:  terminating autovacuum process due to administrator command
worker_2   | LOG:  worker process: Citus Maintenance Daemon: 12407/10 (PID 54) exited with exit code 1
worker_2   | LOG:  shutting down
worker_2   | LOG:  database system is shut down
worker_2   |  done
worker_2   | server stopped

[lfittl]

@once-ler It seems you'll need to remove the volumes and have an empty data directory, since you can't re-use the Postgres 9.6 data directory when running on 10. Whilst there are ways to upgrade, assuming this is your development environment I would just reset it.

The following should do it:

docker-compose -p citus -v down
docker-compose -p citus up

[once-ler]

Thanks @lfittl That did the trick.

When I do a sudo -u postgres psql -h localhost and check the number of workers, I always get 0. Is there something else I need to do?

postgres=# SELECT master_get_active_worker_nodes();
 master_get_active_worker_nodes 
--------------------------------
(0 rows)

[lfittl]

@once-ler I'm wondering if the worker nodes correctly registered - what does docker ps show?

[once-ler]

@lfittl I also tried docker-compose -p citus scale worker=2

docker ps shows the workers are up but the query still returns 0.

CONTAINER ID        IMAGE                   COMMAND                  CREATED             STATUS                    PORTS                    NAMES
1a79cd48785a        citusdata/citus:7.1.1   "docker-entrypoint.sh"   57 minutes ago      Up 57 minutes (healthy)   5432/tcp                 citus_worker_2
247ab7fb6044        citusdata/citus:7.1.1   "docker-entrypoint.sh"   59 minutes ago      Up 59 minutes (healthy)   5432/tcp                 citus_worker_1
66e6b817c2a5        citusdata/citus:7.1.1   "docker-entrypoint.sh"   59 minutes ago      Up 59 minutes (healthy)   0.0.0.0:5432->5432/tcp   citus_master

This is logs from citus_worker_1:

PostgreSQL init process complete; ready for start up.

2017-12-13 21:28:34.283 UTC [1] WARNING:  citus.enable_deadlock_prevention is deprecated and it has no effect. The flag will be removed in the next release.
2017-12-13 21:28:34.289 UTC [1] LOG:  number of prepared transactions has not been configured, overriding
2017-12-13 21:28:34.289 UTC [1] DETAIL:  max_prepared_transactions is now set to 200
2017-12-13 21:28:34.289 UTC [1] LOG:  listening on IPv4 address "0.0.0.0", port 5432
2017-12-13 21:28:34.289 UTC [1] LOG:  listening on IPv6 address "::", port 5432
2017-12-13 21:28:34.381 UTC [1] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
2017-12-13 21:28:34.552 UTC [86] LOG:  database system was shut down at 2017-12-13 21:28:34 UTC
2017-12-13 21:28:34.604 UTC [1] LOG:  database system is ready to accept connections