Two started cluster on the same data directory and port

On 11/30/23 05:37, Loles wrote:

Hi!

I have this situation on a server that is giving me problems.

postgres@hostname:~$ pg_lsclusters
Ver Cluster  Port Status Owner    Data directory              Log file
14  main     5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
14  main_old 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

How can I stop, even delete, the main_old cluster?

Assuming you are on Debian/Ubuntu.

1) sudo pg_ctlcluster stop 14 main_old

This will stop the server for this session

2) sudo vi /etc/postgresql/14/main_old/start.conf

Change auto to manual.
This will keep the server from restarting on reboot.
You can still start it using
sudo pg_ctlcluster start 14 main_old
Though if you want to do that then:
vi /etc/postgresql/14/main_old/postgresql.conf
and change port to something other then 5432.

3) If you have determined you really don't need that server and have all
the data safely backed up or moved to the other server, then(THIS IS NOT
REVERSIBLE):

sudo pg_dropcluster 14 main_old

Thanks!

--
Adrian Klaver
adrian.klaver@aklaver.com

On 11/30/23 07:23, Adrian Klaver wrote:

On 11/30/23 05:37, Loles wrote:

Hi!

I have this situation on a server that is giving me problems.

postgres@hostname:~$ pg_lsclusters
Ver Cluster  Port Status Owner    Data directory              Log file
14  main     5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
14  main_old 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

How can I stop, even delete, the main_old cluster?

Assuming you are on  Debian/Ubuntu.

1) sudo pg_ctlcluster stop 14 main_old

    This will stop the server for this session

2) sudo vi  /etc/postgresql/14/main_old/start.conf

    Change auto to manual.
    This will keep the server from restarting on reboot.
    You can still start it using
     sudo pg_ctlcluster start 14 main_old
    Though if you want to do that then:
        vi /etc/postgresql/14/main_old/postgresql.conf
     and change port to something other then 5432.

3) If you have determined you really don't need that server and have all
the data safely backed up or moved to the other server, then(THIS IS NOT
REVERSIBLE):

    sudo pg_dropcluster 14 main_old

Blaming it on not enough coffee, but I just noticed both clusters are
pointing at the same data directory. I would NOT do:

sudo pg_dropcluster 14 main_old

Thanks!

--
Adrian Klaver
adrian.klaver@aklaver.com

Every time I stop the main_old cluster with sudo pg_ctlcluster stop 14
main_old, both are stopped. And when I raise sudo pg_ctlcluster start 14
main, both are raised.

I don't know how I got into this situation and I don't know how to resolve
it.

El jue, 30 nov 2023 a las 17:34, Loles (<lolesft@gmail.com>) escribió:

On 11/30/23 08:34, Loles wrote:

Please reply to list also.
Ccing list.

Indeed, that is the problem, they are two different clusters, pointing
to the same data directory and reading from the same port.

I can't change the port because it is the same postgresql.conf

I don't know how to disassociate the main_old cluster from the data
directory or how to resolve this situation.

El jue, 30 nov 2023 a las 16:32, Adrian Klaver
(<adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>>) escribió:

On 11/30/23 07:23, Adrian Klaver wrote:

On 11/30/23 05:37, Loles wrote:

Hi!

I have this situation on a server that is giving me problems.

postgres@hostname:~$ pg_lsclusters
Ver Cluster  Port Status Owner    Data directory

 Log file

14  main     5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
14  main_old 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

How can I stop, even delete, the main_old cluster?

Assuming you are on  Debian/Ubuntu.

1) sudo pg_ctlcluster stop 14 main_old

      This will stop the server for this session

2) sudo vi  /etc/postgresql/14/main_old/start.conf

      Change auto to manual.
      This will keep the server from restarting on reboot.
      You can still start it using
       sudo pg_ctlcluster start 14 main_old
      Though if you want to do that then:
          vi /etc/postgresql/14/main_old/postgresql.conf
       and change port to something other then 5432.

3) If you have determined you really don't need that server and

have all

the data safely backed up or moved to the other server, then(THIS

IS NOT

REVERSIBLE):

      sudo pg_dropcluster 14 main_old

Blaming it on not enough coffee, but I just noticed both clusters are
pointing at the same data directory. I would NOT do:

     sudo pg_dropcluster 14 main_old

Thanks!

--
Adrian Klaver
adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>

--
Adrian Klaver
adrian.klaver@aklaver.com

On 11/30/23 08:37, Loles wrote:

Every time I stop the main_old cluster with sudo pg_ctlcluster stop 14
main_old, both are stopped. And when I raise sudo pg_ctlcluster start 14
main, both are raised.

I don't know how I got into this situation and I don't know how to
resolve it.

Is there both?:

/etc/postgresql/14/main/
/etc/postgresql/14/main_old/

El jue, 30 nov 2023 a las 17:34, Loles (<lolesft@gmail.com
<mailto:lolesft@gmail.com>>) escribió:

Indeed, that is the problem, they are two different clusters,
pointing to the same data directory and reading from the same port.

I can't change the port because it is the same postgresql.conf

I don't know how to disassociate the main_old cluster from the data
directory or how to resolve this situation.

El jue, 30 nov 2023 a las 16:32, Adrian Klaver
(<adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>>)
escribió:

On 11/30/23 07:23, Adrian Klaver wrote:

On 11/30/23 05:37, Loles wrote:

Hi!

I have this situation on a server that is giving me problems.

postgres@hostname:~$ pg_lsclusters
Ver Cluster  Port Status Owner    Data directory

 Log file

14  main     5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
14  main_old 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

How can I stop, even delete, the main_old cluster?

Assuming you are on  Debian/Ubuntu.

1) sudo pg_ctlcluster stop 14 main_old

      This will stop the server for this session

2) sudo vi  /etc/postgresql/14/main_old/start.conf

      Change auto to manual.
      This will keep the server from restarting on reboot.
      You can still start it using
       sudo pg_ctlcluster start 14 main_old
      Though if you want to do that then:
          vi /etc/postgresql/14/main_old/postgresql.conf
       and change port to something other then 5432.

3) If you have determined you really don't need that server

and have all

the data safely backed up or moved to the other server,

then(THIS IS NOT

REVERSIBLE):

      sudo pg_dropcluster 14 main_old

Blaming it on not enough coffee, but I just noticed both
clusters are
pointing at the same data directory. I would NOT do:

     sudo pg_dropcluster 14 main_old

Thanks!

--
Adrian Klaver
adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>

--
Adrian Klaver
adrian.klaver@aklaver.com

Yes, it exists but it is obsolete. The files and directories are all dated
2022.

It was obsolete and no longer valid, it could be deleted. Only the main
cluster is correct.

El jue, 30 nov 2023 a las 17:43, Adrian Klaver (<adrian.klaver@aklaver.com>)
escribió:

On 11/30/23 08:46, Loles wrote:

Yes, it exists but it is obsolete. The files and directories are all
dated 2022.

It was obsolete and no longer valid, it could be deleted. Only the main
cluster is correct.

First try:

vi /etc/postgresql/14/main_old/start.conf

and change auto to disabled.

Stop the server(s) and then:

sudo pg_ctlcluster starp 14 main

El jue, 30 nov 2023 a las 17:43, Adrian Klaver
(<adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>>) escribió:

On 11/30/23 08:37, Loles wrote:

Every time I stop the main_old cluster with sudo pg_ctlcluster

stop 14

main_old, both are stopped. And when I raise sudo pg_ctlcluster

start 14

main, both are raised.

I don't know how I got into this situation and I don't know how to
resolve it.

Is there both?:

    /etc/postgresql/14/main/
    /etc/postgresql/14/main_old/

El jue, 30 nov 2023 a las 17:34, Loles (<lolesft@gmail.com

<mailto:lolesft@gmail.com>

<mailto:lolesft@gmail.com <mailto:lolesft@gmail.com>>>) escribió:

     Indeed, that is the problem, they are two different clusters,
     pointing to the same data directory and reading from the same

port.

     I can't change the port because it is the same postgresql.conf

     I don't know how to disassociate the main_old cluster from

the data

     directory or how to resolve this situation.

     El jue, 30 nov 2023 a las 16:32, Adrian Klaver
     (<adrian.klaver@aklaver.com

<mailto:adrian.klaver@aklaver.com> <mailto:adrian.klaver@aklaver.com
<mailto:adrian.klaver@aklaver.com>>>)

     escribió:

         On 11/30/23 07:23, Adrian Klaver wrote:
          > On 11/30/23 05:37, Loles wrote:
          >> Hi!
          >>
          >> I have this situation on a server that is giving me

problems.

          >>
          >> postgres@hostname:~$ pg_lsclusters
          >> Ver Cluster  Port Status Owner    Data directory
           Log file
          >> 14  main     5432 online postgres

/var/lib/postgresql/14/main

          >> /var/log/postgresql/postgresql-14-main.log
          >> 14  main_old 5432 online postgres

/var/lib/postgresql/14/main

          >> /var/log/postgresql/postgresql-14-main_old.log
          >>
          >> How can I stop, even delete, the main_old cluster?
          >
          > Assuming you are on  Debian/Ubuntu.
          >
          > 1) sudo pg_ctlcluster stop 14 main_old
          >
          >      This will stop the server for this session
          >
          > 2) sudo vi  /etc/postgresql/14/main_old/start.conf
          >
          >      Change auto to manual.
          >      This will keep the server from restarting on reboot.
          >      You can still start it using
          >       sudo pg_ctlcluster start 14 main_old
          >      Though if you want to do that then:
          >          vi /etc/postgresql/14/main_old/postgresql.conf
          >       and change port to something other then 5432.
          >
          > 3) If you have determined you really don't need that

server

         and have all
          > the data safely backed up or moved to the other server,
         then(THIS IS NOT
          > REVERSIBLE):
          >
          >      sudo pg_dropcluster 14 main_old

         Blaming it on not enough coffee, but I just noticed both
         clusters are
         pointing at the same data directory. I would NOT do:

               sudo pg_dropcluster 14 main_old

          >
          >>
          >> Thanks!
          >>
          >

         --
         Adrian Klaver
adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>

<mailto:adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>>

--
Adrian Klaver
adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>

--
Adrian Klaver
adrian.klaver@aklaver.com

root@hostname:~# su - postgres
postgres@hostname:~$ *nano /etc/postgresql/14/main_old/start.conf*
postgres@hostname:~$ exit
cerrar sesión
root@hostname:~# *pg_ctlcluster stop 14 main*
*(change auto for disabled and exit with save)*
root@hostname:~# ps -ef | grep postgres
root 2609 1512 0 18:24 pts/1 00:00:00 grep postgres
root@hostname:~# pg_lsclusters
Ver Cluster Port Status Owner Data directory Log file
*14 main 5432 down* postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
*14 main_old 5432 down * postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log
root@hostname:~#
root@hostname:~#
root@hostname:~# *pg_ctlcluster start 14 main*
root@hostname:~# ps -ef | grep postgres
postgres 2620 1 2 18:24 ? 00:00:00
/usr/lib/postgresql/14/bin/postgres -D /var/lib/postgresql/14/main -c
config_file=/etc/postgresql/14/main/postgresql.conf
postgres 2625 2620 0 18:24 ? 00:00:00 postgres: 14/main:
checkpointer
postgres 2626 2620 0 18:24 ? 00:00:00 postgres: 14/main:
background writer
postgres 2627 2620 0 18:24 ? 00:00:00 postgres: 14/main:
walwriter
postgres 2628 2620 0 18:24 ? 00:00:00 postgres: 14/main:
autovacuum launcher
postgres 2629 2620 0 18:24 ? 00:00:00 postgres: 14/main:
stats collector
postgres 2630 2620 0 18:24 ? 00:00:00 postgres: 14/main:
logical replication launcher
root 2642 1512 0 18:24 pts/1 00:00:00 grep postgres
root@hostname:~# pg_lsclusters
Ver Cluster Port Status Owner Data directory Log file
*14 main 5432 online* postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
*14 main_old 5432 online *postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

Start, up the two cluster.

El jue, 30 nov 2023 a las 18:01, Adrian Klaver (<adrian.klaver@aklaver.com>)
escribió:

On 11/30/23 09:27, Loles wrote:

root@hostname:~# pg_lsclusters
Ver Cluster  Port Status Owner    Data directory              Log file
*14  main     5432 online* postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
*14  main_old 5432 online *postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

Start, up the two cluster.

Before going any further make sure you have a current backup of the
database.

Hmm, are /etc/postgresql/14/main/ and /etc/postgresql/14/main_old/
symlinked together?

If not what if you move the ~14/main_old/ directory somewhere else and
then restart Postgres?

Adrian Klaver
adrian.klaver@aklaver.com

No, they are not symbolic links. I have thought about moving the main_old
directory somewhere else and it doesn't change anything because the cluster
main_old is linked to the main data directory. In fact, neither the log nor
any main_old data file changes, they are all from 2020. It's a strange case

El jue., 30 nov. 2023 16:23, Adrian Klaver <adrian.klaver@aklaver.com>
escribió:

On 11/30/23 10:35, Adrian Klaver wrote:

On 11/30/23 09:27, Loles wrote:

root@hostname:~# pg_lsclusters
Ver Cluster  Port Status Owner    Data directory              Log file
*14  main     5432 online* postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
*14  main_old 5432 online *postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

Start, up the two cluster.

Before going any further make sure you have a current backup of the
database.

Hmm, are /etc/postgresql/14/main/ and /etc/postgresql/14/main_old/
symlinked together?

If not what if you move the ~14/main_old/ directory somewhere else and
then restart Postgres?

Alright results of testing here:

tree /etc/postgresql/14
/etc/postgresql/14
├── main
│   ├── conf.d
│   ├── environment
│   ├── pg_ctl.conf
│   ├── pg_hba.conf
│   ├── pg_ident.conf
│   ├── postgresql.conf
│   └── start.conf
└── main_old
├── conf.d
├── environment
├── pg_ctl.conf
├── pg_hba.conf
├── pg_ident.conf
├── postgresql.conf
└── start.conf

Where I just copied main/ to main_old/.

sudo pg_ctlcluster start 14 main

pg_lsclusters
Ver Cluster Port Status Owner Data directory Log file
14 main 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
14 main_old 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

sudo pg_ctlcluster stop 14 main

pg_lsclusters
Ver Cluster Port Status Owner Data directory Log file
14 main 5432 down postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log
14 main_old 5432 down postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main_old.log

cd /etc/postgresql/14/
sudo rm -r main_old/
or
sudo cp -r main_old <some_new_path>

sudo pg_ctlcluster start 14 main

pg_lsclusters
Ver Cluster Port Status Owner Data directory Log file
14 main 5432 online postgres /var/lib/postgresql/14/main
/var/log/postgresql/postgresql-14-main.log

I would need to understand systemd better to be sure, but I'm thinking
the duplicated cluster are happening here:

/lib/systemd/system-generators/postgresql-generator

#!/bin/sh

# This systemd generator creates dependency symlinks that make all
PostgreSQL
# clusters with "auto" in their start.conf file be started/stopped/reloaded
# when postgresql.service is started/stopped/reloaded.

set -eu

gendir="$1"
wantdir="$1/postgresql.service.wants"
bindir="/usr/lib/postgresql/"
#redhat# bindir="/usr/pgsql-"
pgservice="/lib/systemd/system/postgresql@.service"

mkdir -p "$wantdir"

for conf in /etc/postgresql/*/*/postgresql.conf; do
# abort loop if glob was not expanded (but accept dead symlinks)
if ! test -e "$conf" && ! test -L "$conf"; then continue; fi

dir="${conf%/*}"

# evaluate start.conf
if [ -e "$dir/start.conf" ]; then
start=$(sed 's/#.*$//; /^[[:space:]]*$/d; s/^\s*//;
s/\s*$//' "$dir/start.conf")
else
start=auto
fi
[ "$start" = "auto" ] || continue

verdir="${dir%/*}"
version="${verdir##*/}"
test -x "$bindir$version/bin/postgres" || continue # package
got removed
cluster="${dir##*/}"
ln -s "$pgservice" "$wantdir/postgresql@$version-$cluster.service"
done

exit 0

Adrian Klaver
adrian.klaver@aklaver.com

--
Adrian Klaver
adrian.klaver@aklaver.com

On 11/30/23 12:35, Adrian Klaver wrote:

On 11/30/23 10:35, Adrian Klaver wrote:

On 11/30/23 09:27, Loles wrote:

cd /etc/postgresql/14/
sudo rm -r main_old/
  or
sudo cp -r main_old <some_new_path>

Arrgh.

sudo mv -r main_old <some_new_path>

Memo to self don't eat lunch and copy/paste at same time.

--
Adrian Klaver
adrian.klaver@aklaver.com

El día jueves, noviembre 30, 2023 a las 02:10:25p. m. -0800, Adrian Klaver escribió:

On 11/30/23 12:35, Adrian Klaver wrote:

On 11/30/23 10:35, Adrian Klaver wrote:

On 11/30/23 09:27, Loles wrote:

cd /etc/postgresql/14/
sudo rm -r main_old/
  or
sudo cp -r main_old <some_new_path>

Arrgh.

sudo mv -r main_old <some_new_path>

Memo to self don't eat lunch and copy/paste at same time.

Hmmm

purism@pureos:~$ uname -s
Linux
purism@pureos:~$ mv -r foo bar
mv: invalid option -- 'r'
Try 'mv --help' for more information.

matthias
--
Matthias Apitz, ✉ guru@unixarea.de, http://www.unixarea.de/ +49-176-38902045
Public GnuPG key: http://www.unixarea.de/key.pub

I am not at war with Russia.
Я не воюю с Россией.
Ich bin nicht im Krieg mit Russland.

On 11/30/23 20:43, Matthias Apitz wrote:

El día jueves, noviembre 30, 2023 a las 02:10:25p. m. -0800, Adrian Klaver escribió:

On 11/30/23 12:35, Adrian Klaver wrote:

On 11/30/23 10:35, Adrian Klaver wrote:

On 11/30/23 09:27, Loles wrote:

cd /etc/postgresql/14/
sudo rm -r main_old/
  or
sudo cp -r main_old <some_new_path>

Arrgh.

sudo mv -r main_old <some_new_path>

Memo to self don't eat lunch and copy/paste at same time.

Hmmm

purism@pureos:~$ uname -s
Linux
purism@pureos:~$ mv -r foo bar
mv: invalid option -- 'r'

The gift that keeps on giving. Third time's a charm?:

sudo mv main_old/ <some_new_path>

Try 'mv --help' for more information.

matthias

--
Adrian Klaver
adrian.klaver@aklaver.com

Yes, it seems to be caused by systemd.

I'll investigate there. If not, I can upgrade PostgreSQL and get rid of
server 14.

Thank you very much Adrian... I'm going to get a cup of coffee because the
topic requires it :D

El vie, 1 dic 2023 a las 6:19, Adrian Klaver (<adrian.klaver@aklaver.com>)
escribió:

On 12/1/23 01:17, Loles wrote:

Yes, it seems to be caused by systemd.

To be more accurate it is caused by the the file postgresql-generator
located here:

/lib/systemd/system-generators/

That file is installed by the Postgres Debian/Ubuntu package, so it is
the Postgres install that is using systemd in this manner.

I'll investigate there. If not, I can upgrade PostgreSQL and get rid of
server 14.

If your directory structure resembles:

/etc/postgresql/14
├── main
│ ├── conf.d
│ ├── environment
│ ├── pg_ctl.conf
│ ├── pg_hba.conf
│ ├── pg_ident.conf
│ ├── postgresql.conf
│ └── start.conf
└── main_old
├── conf.d
├── environment
├── pg_ctl.conf
├── pg_hba.conf
├── pg_ident.conf
├── postgresql.conf
└── start.conf

Then you just need to:

1) Stop the server
2) mv or rm the main_old/ directory
3) Restart the server

Thank you very much Adrian... I'm going to get a cup of coffee because
the topic requires it :D

El vie, 1 dic 2023 a las 6:19, Adrian Klaver (<adrian.klaver@aklaver.com
<mailto:adrian.klaver@aklaver.com>>) escribió:

On 11/30/23 20:43, Matthias Apitz wrote:

El día jueves, noviembre 30, 2023 a las 02:10:25p. m. -0800,

Adrian Klaver escribió:

On 11/30/23 12:35, Adrian Klaver wrote:

On 11/30/23 10:35, Adrian Klaver wrote:

On 11/30/23 09:27, Loles wrote:

cd /etc/postgresql/14/
sudo rm -r main_old/
     or
sudo cp -r main_old <some_new_path>

Arrgh.

      sudo mv -r main_old <some_new_path>

Memo to self don't eat lunch and copy/paste at same time.

Hmmm

purism@pureos:~$ uname -s
Linux
purism@pureos:~$ mv -r foo bar
mv: invalid option -- 'r'

The gift that keeps on giving. Third time's a charm?:

        sudo mv  main_old/ <some_new_path>

Try 'mv --help' for more information.

       matthias

--
Adrian Klaver
adrian.klaver@aklaver.com <mailto:adrian.klaver@aklaver.com>

--
Adrian Klaver
adrian.klaver@aklaver.com