Docker compose - not working - keeps restarting container. about traccar-docker

[magna-z]

After running my docker compose, the container keeps restarting. Here is my docker-compose file.
Am i doing something incorrect. Please advice.

version: '3'
services:
  traccar:
    image: "traccar/traccar:latest"
    restart: always
    entrypoint:
      - java
      - -Xms2g
      - -Xmx2g
      - -Djava.net.preferIPv4Stack=true    
    ports:
      - "5000-5150:5000-5150"
      - "5000-5150:5000-5150/udp"
      - "8082:8082"
    volumes:
      - "/data/traccar/logs:/opt/traccar/logs:rw"
      - "/data/traccar/data:/opt/traccar/data:rw"
      - "/root/traccar/traccar.xml:/opt/traccar/conf/traccar.xml:ro"
      - "/etc/localtime:/etc/localtime:ro"
      - "/etc/timezone:/etc/timezone:ro"
    ulimits:
      nproc: 65535
      nofile:
        soft: 50000
        hard: 50000      

In Docker, when you run container with custom ENTRYPOINT, original CMD in image are removed. See docker inspect traccar after create container.
If you need to change ENTRYPOINT for container in docker-compose, use:

...
    entrypoint:
    - java
    - -Xms2g
    - -Xmx2g
    - -Djava.net.preferIPv4Stack=true
    command:
    - -jar
    - tracker-server.jar
    - conf/traccar.xml
...

or, set start service from entrypoint only

...
    entrypoint:
    - java
    - -Xms2g
    - -Xmx2g
    - -Djava.net.preferIPv4Stack=true
    - -jar
    - tracker-server.jar
    - conf/traccar.xml
...

For run docker container in raw command:
docker run ... --entrypoint java traccar/traccar:4.1 -Xms2g -Xmx2g -Djava.net.preferIPv4Stack=true -jar tracker-server.jar conf/traccar.xml

This is feature/bug of using java with options in docker and i will be search the best realization of this.

[LeoRX]

not sure if its related. I'm having the same issue when try to use mysqldb and result in this error.

Exception in thread "main" java.lang.RuntimeException: java.lang.ArrayIndexOutOfBoundsException: 1,
	at org.traccar.Main.run(Main.java:152),
	at org.traccar.Main.main(Main.java:104),
Caused by: java.lang.ArrayIndexOutOfBoundsException: 1,
	at liquibase.datatype.DataTypeFactory.fromDescription(DataTypeFactory.java:251),
	at liquibase.change.core.CreateTableChange.generateStatements(CreateTableChange.java:70),
	at liquibase.change.AbstractChange.generateStatementsVolatile(AbstractChange.java:287),
	at liquibase.change.AbstractChange.warn(AbstractChange.java:358),
	at liquibase.changelog.visitor.ValidatingVisitor.visit(ValidatingVisitor.java:109),
	at liquibase.changelog.ChangeLogIterator.run(ChangeLogIterator.java:78),
	at liquibase.changelog.DatabaseChangeLog.validate(DatabaseChangeLog.java:269),
	at liquibase.Liquibase.update(Liquibase.java:198),
	at liquibase.Liquibase.update(Liquibase.java:179),
	at liquibase.Liquibase.update(Liquibase.java:175),
	at org.traccar.database.DataManager.initDatabaseSchema(DataManager.java:306),
	at org.traccar.database.DataManager.<init>(DataManager.java:91),
	at org.traccar.Context.init(Context.java:382),
	at org.traccar.Main.run(Main.java:110),
	... 1 more,

https://www.traccar.org/forums/topic/another-40-greater41-problem/ has more detail.

in the meantime, use default h2db or use version 4 instead.

[delatitude]

I must have mentioned a few more details.

Environment

Type: VM
Virtualization: microsoft
Operating System: Ubuntu 16.04.3 LTS
Kernel: Linux 4.15.0-1025-azure
Architecture: x86-64
CPU. 8 cores. Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
RAM. 28 GB

1. Issue is with setting jvm to a higher value (eg. Xms2g Xmx2g).
2. If i run with its default jvm settings (Xms512m ,Xmx512m.), then it works perfectly for a while untill the memory fills up.
3. I am using its default database h2db.
4. Now i've rolled back to an old version but still this new version doesnt work.

[delatitude]

This was good. It works now without restarting. I will keep monitoring and get back here for confirmation.

However, the logs aren't being produced. If i run command
"docker-compose logs -f" it returns "Error grabbing logs: EOF"... any ideas please?

[magna-z]

This was good. It works now without restarting. I will keep monitoring and get back here for confirmation.

However, the logs aren't being produced. If i run command
"docker-compose logs -f" it returns "Error grabbing logs: EOF"... any ideas please?

For "Error grabbing logs: EOF" can may be many reasons. By closed issue i recommend update docker and docker-compose on latest versions. What versions now?
May be you change log-driver in /etc/docker/daemon.json and docker-compose have not access on this log storage?