Docker Volume - PG::DiskFull - No space left on device

Question

How to solve a Disk full error in Docker Volume using a Postgresql database.

I'm not sure if the issue is with docker, docker volumes or Postgres configuration.

Environment

Docker on MAC - Postgresql

The following error happens while running a Rails application.

PG::DiskFull: ERROR: could not write to file "base/pgsql_tmp/pgsql_tmp125.6.sharedfileset/i35of64.p0.0": No space left on device

I'm also having these types of issues when importing database dump

ERROR: could not extend file "base/1051062/1259": No space left on device

HINT: Check free disk space.

STATEMENT: CREATE INDEX index_some_table_key ON public.some_table USING btree (xyz_id, abc_key) WHERE ((NOT deleted) AND (abc_key IS NOT NULL));

Container shell disk size

docker exec -it pg13 bash

root@51fe1c96f701:/# df -h
Filesystem      Size  Used Avail Use% Mounted on
overlay          59G   56G   26M 100% /
tmpfs            64M     0   64M   0% /dev
tmpfs           2.0G     0  2.0G   0% /sys/fs/cgroup
shm              64M  1.3M   63M   2% /dev/shm
/dev/vda1        59G   56G   26M 100% /etc/hosts
tmpfs           2.0G     0  2.0G   0% /proc/acpi
tmpfs           2.0G     0  2.0G   0% /sys/firmware

I have been gradually working through different ways of using pg_restore

# many indexes in the database fail to create
pg_restore -h 127.0.0.1 -U postgres -d "mydb" "mydump.dump"

# when running data only, the restore works
pg_restore --section=pre-data --section=data -h 127.0.0.1 -U postgres -d "mydb" "mydump.dump"

# when running for specific TOCs, I sometimes have success and sometimes have out of disk space errors
pg_restore -v -L x.txt -h 127.0.0.1 -U postgres -d "printspeak_development" "salescrm.dump"

# contents of x.txt
;5971; 1259 31041251 INDEX public index_account_history_data_tenant_source_account_platform_id salescrm
;6138; 1259 2574937 INDEX public index_action_logs_on_action salescrm

Database that was restored

I have a created a data base and restored a dump from production.

.dump file was 8GB in size and the imported database is 29GB in size

SELECT pg_size_pretty( pg_database_size('mydb_development') );

 pg_size_pretty
----------------
 29 GB

Docker Compose

volumes:        
   pgdata:

services:
  pg13:
    container_name: "pg13"
    image: "postgres:latest"
    environment:
      POSTGRES_USER: "postgres"
      POSTGRES_PASSWORD: "xyz"

    # Increase the Write Ahead Log size because of warnings.
    command: "postgres -c max_wal_size=2GB"

    ports:
      - "5432:5432"
    volumes:
      - pgdata:/var/lib/postgresql/data

What made this issue difficult to figure out (misled me).

I had enough disk space in the Docker for the Database + Data + Many of the indexes.

But not all the indexes and there was also some randomness to the create indexes, sometimes they worked, sometimes they didn't.

Answer 1

The issue was with the amount of disk space being allocated by Docker and ended up being a simple fix in the Docker GUI.