Re: Issues shutting down PostgreSQL 10 cleanly

Hi Laurenz,

/etc/init.d/postgresql stop prompts the same condition.

Simon - I get that ...

Apparently the solution is to run a checkpoint first and then shutdown?
http://rhaas.blogspot.com/2015/03/postgresql-shutdown.html

I'll try this likely tomorrow and let you know.

--Stephen

-----Original Message-----
From: Laurenz Albe [mailto:laurenz(dot)albe(at)cybertec(dot)at]
Sent: Tuesday, October 24, 2017 12:20 AM
To: Stephen Froehlich <s(dot)froehlich(at)cablelabs(dot)com>; pgsql-novice(at)postgresql(dot)org
Subject: Re: [NOVICE] Issues shutting down PostgreSQL 10 cleanly

CableLabs WARNING: The sender of this email could not be validated and may not match the person in the "From" field.

Stephen Froehlich wrote:
> I am in the midst of heavy, large write operations on a new database this is currently around 23 billion lines.
>
> My max_wal_size is 256 GB and my checkpoint_timeout is 12 hours. (Smaller values were hindering the writes.) Note: this is on Ubuntu 16.04 LTS.
>
> I am having trouble shutting down postgres cleanly – how can I fix that?
>
> A “/etc/init.d/postgresql restart” command gave the following log entries:
>
> 2017-10-20 09:17:55.069 MDT [16388] LOG: received fast shutdown
> request
> 2017-10-20 09:17:55.069 MDT [16388] LOG: aborting any active
> transactions
> 2017-10-20 09:17:55.069 MDT [20336] FATAL: terminating autovacuum
> process due to administrator command
> 2017-10-20 09:17:55.069 MDT [20540] FATAL: terminating autovacuum
> process due to administrator command
> 2017-10-20 09:17:55.069 MDT [20132] FATAL: terminating autovacuum
> process due to administrator command
> 2017-10-20 09:17:55.070 MDT [1486] sfroehlich(at)mediacom_df_vd FATAL:
> terminating connection due to administrator command
> 2017-10-20 09:17:55.070 MDT [1488] sfroehlich(at)mediacom_df_vd FATAL:
> terminating connection due to administrator command
> 2017-10-20 09:17:55.070 MDT [1487] sfroehlich(at)mediacom_df_vd FATAL:
> terminating connection due to administrator command
> 2017-10-20 09:17:55.070 MDT [1489] sfroehlich(at)mediacom_df_vd FATAL:
> terminating connection due to administrator command
> 2017-10-20 09:17:55.070 MDT [1485] sfroehlich(at)mediacom_df_vd FATAL:
> terminating connection due to administrator command
> 2017-10-20 09:17:55.074 MDT [16388] LOG: worker process: logical
> replication launcher (PID 16400) exited with exit code 1
> 2017-10-20 09:19:08.363 MDT [16395] LOG: shutting down
> 2017-10-20 09:20:29.559 MDT [32352] LOG: listening on IPv4 address
> "0.0.0.0", port 5432
> 2017-10-20 09:20:29.559 MDT [32352] LOG: listening on IPv6 address
> "::", port 5432
> 2017-10-20 09:20:29.560 MDT [32352] LOG: listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
> 2017-10-20 09:20:57.956 MDT [32454] LOG: database system shutdown was
> interrupted; last known up at 2017-10-20 09:19:08 MDT
> 2017-10-20 09:21:03.598 MDT [32454] LOG: database system was not
> properly shut down; automatic recovery in progress
> 2017-10-20 09:21:03.646 MDT [32454] LOG: redo starts at 972/DB0B0838
> pg_ctl: server did not start in time
> 2017-10-20 09:21:29.645 MDT [32352] LOG: received smart shutdown
> request
> 2017-10-20 09:21:29.735 MDT [32558] LOG: shutting down
> 2017-10-20 09:21:30.019 MDT [32352] LOG: database system is shut down

What happens if you run /etc/init.d/postgresql stop?

Clean shutdown? Any processes left running?

I would examine /etc/init.d/postgresql and see if it does evil things like porcibly remove the postmaster.pid file.

Yours,
Laurenz Albe