Problem
When trying to re-initialise MongoDB (IE: perform a clean wipe), it seems that the docker-entrypoint.sh is failing to detect a 'uninitialised' instance of Mongo.
Some times when I execute the following steps I do see the following message within the mongo container logs.
MongoDB init process complete; ready for start up.
Other times, it does not print the message above and therefore does not initialise.
Steps to reproduce
Steps to run
docker volume rm mongodb-data
docker volume rm mongodb_configdb_data
docker volume create mongodb-data
docker volume create mongodb_configdb_data
# When this runs, it shows that the `/data/db` directory is EMPTY
docker run --rm \
-v mongodb-data:/data/db:Z \
-v mongodb_configdb_data:/data/configdb:Z \
alpine ls -R /data/db
docker compose up -d -V mongo
docker container logs -f mongo
# NOTE: in the logs, the container has initialised successfully IF you see the following message
# "MongoDB init process complete; ready for start up."
Reference: docker-compose.yml
# other services...
services:
mongo:
image: mongo:5.0.8
networks:
- myNetwork
volumes:
- mongodb-data:/data/db:Z
- mongodb_configdb_data:/data/configdb:Z
environment:
MONGO_INITDB_ROOT_USERNAME: someUsername
MONGO_INITDB_ROOT_PASSWORD: somePassword
MONGO_INITDB_DATABASE: someDbName
entrypoint:
- bash
- -c
- | # NOTE: This is coming from a Docker secret
cp /usr/local/data/replica.key /usr/local/data/replication.container.key
chmod 400 /usr/local/data/replication.container.key
chown 999:999 /usr/local/data/replication.container.key
exec docker-entrypoint.sh $$@
command: "--bind_ip_all --replSet rs0 --keyFile /usr/local/data/replication.container.key"
healthcheck:
test: echo 'db.runCommand("ping").ok' | mongosh localhost:27017/${MONGO_DB_NAME} --quiet
interval: 10s
timeout: 30s
retries: 10
Other info
My Docker volumes are NFS mounted volumes running in AWS EFS. When the MongoDB container starts and does not initialise - I see my EFS volume throughput spike and become pinned at 100% until I kill the now-stuck MongoDB container.
Image tag: mongo:5.0.8
Docker version: v25.0.1
Docker Compose version: v2.24.2
docker inspect mongodb-data
[
{
"CreatedAt": "2024-03-15T06:39:47Z",
"Driver": "local",
"Labels": null,
"Mountpoint": "/var/lib/docker/volumes/mongodb-data/_data",
"Name": "mongodb-data",
"Options": {
"device": ":/mongodb-data",
"o": "nfsvers=4.1,rsize=1048576,wsize=1048576,bg,hard,nolock,noatime,nointr,timeo=600,retrans=2,noresvport,addr=fs-redacted.efs.ap-southeast-2.amazonaws.com,rw",
"type": "nfs"
},
"Scope": "local"
}
]
Problem
When trying to re-initialise MongoDB (IE: perform a clean wipe), it seems that the
docker-entrypoint.shis failing to detect a 'uninitialised' instance of Mongo.Some times when I execute the following steps I do see the following message within the
mongocontainer logs.Other times, it does not print the message above and therefore does not initialise.
Steps to reproduce
Steps to run
Reference: docker-compose.yml
Other info
My Docker volumes are NFS mounted volumes running in AWS EFS. When the MongoDB container starts and does not initialise - I see my EFS volume throughput spike and become pinned at 100% until I kill the now-stuck MongoDB container.
Image tag:
mongo:5.0.8Docker version:
v25.0.1Docker Compose version:
v2.24.2docker inspect mongodb-data