could not create shared memory segment: No space left on device

[EugenKon]

perl -d -Ilib -Ilocal/lib/perl5 -It/lib t/module/db/postgres_rls.t

2025-07-05 17:52:56.843 EDT [15333] FATAL:  could not create shared memory segment: No space left on device
2025-07-05 17:52:56.843 EDT [15333] DETAIL:  Failed system call was shmget(key=37438435, size=56, 03600).
2025-07-05 17:52:56.843 EDT [15333] HINT:  This error does *not* mean that you have run out of disk space.  It occurs either if all available shared memory IDs have been taken, in which case you need to raise the SHMMNI parameter in your kernel, or because the system's overall limit for shared memory has been reached.
	The PostgreSQL documentation contains more information about shared memory configuration.
child process exited with exit code 1
initdb: removing data directory "/var/folders/0c/4gm1slrx1_xf507cswm6ykmw0000gn/T/pgtest.JbXo6/data"
pg_ctl: database system initialization failed
Failed to start postgres after trying 10 potential ports: Failed to open /var/folders/0c/4gm1slrx1_xf507cswm6ykmw0000gn/T/pgtest.JbXo6/data/postmaster.pid: No such file or directory at local/lib/perl5/Test/PostgreSQL.pm line 484.

ipcs -m
IPC status from <running system> as of Sat Jul  5 17:57:18 EDT 2025
T     ID     KEY        MODE       OWNER    GROUP
Shared Memory:
m  65536 0x510f0355 --rw-------   ek-mac    staff
m 327681 0x023a0ad2 --rw-------   ek-mac    staff
m 327682 0x023a2872 --rw-------   ek-mac    staff
m 983043 0x023a39c7 --rw-------   ek-mac    staff
m 327684 0x023a3f21 --rw-------   ek-mac    staff
m 327685 0x023a496f --rw-------   ek-mac    staff
....

ChatGPT says that this could be issue of this module: it does not cleanup properly. Because I run my test under debugger I often CTRL+C. Could be there a way to cleanup everything properly?

[TJC]

This module, Test::PostgreSQL, doesn't manage any shared memory segments directly. The Postgres daemon does, but it cleans that up correctly if terminated properly.

I wonder if your Ctrl-C-ing out of the debugger is preventing some clean-up hooks from being run. Are there any residual postgres daemon processes still running?

[EugenKon]

I do not have PostgreSQL installed on my local machine. It is inside container:

docker ps
CONTAINER ID   IMAGE         COMMAND                  CREATED       STATUS      PORTS                    NAMES
31e9b30a5aee   postgres:17   "docker-entrypoint.s…"   4 weeks ago   Up 3 days   0.0.0.0:5432->5432/tcp   postgres-db-inv

[TJC]

I wonder if the docker container is limiting the amount of shared memory available, but indicating a higher amount to postgres? Or some error along those lines.

Could you give us some info about your operating system and version? eg. "Ubuntu 24.04" or "MacOS 15.6"
If you're on Mac, then which container system are you using? eg. Docker Desktop, Orbstack, etc

[EugenKon]

MacOS 15.6

docker --version
Docker version 27.4.1, build b9d17eaebb
docker-desktop 4.37.2,179585 (upgraded today to 4.43.2,199162 via homebrew )

The issue happens only once a week if I do not reboot my macbook during this period.

Postgres container was build as:

        docker run --name ${DOCKER_CONTAINER} -d -p ${DB_PORT}:5432      \
                -v ${APP_ROOT}/db/pgdata:/var/lib/postgresql/data/pgdata     \
                -e PGDATA=/var/lib/postgresql/data/pgdata                    \
                -e PGTZ=${PGTZ}                                              \
                -e POSTGRES_PASSWORD=${DB_ROOT}                              \
                --shm-size=512m                                              \
                --restart='always'                                           \
                postgres:${PG_VERSION}