spine directory and crontab empty after docker-compose down and restoring

[kevburkett]

Using the cacti_single_install.yml docker-compose example. When bringing down/removing the containers with docker-compose down and restoring with docker-compose up -d the /spine/ directory is empty and original contents of the /etc/crontab is lost. I don't see any errors in logs so I wont post unless you need it.

Example:
After creating new container with persistent volumes for /cacti and /var/lib/mysql

Before snapshot: /spine directory and /etc/crontab has content.

kevin@p-serv-01:~/docker/cacti-test$ docker exec -it cacti-test_cacti_1 /bin/bash
[root@b916fb282470 /]# ls /spine/
bin  etc  share
[root@b916fb282470 /]# cat /etc/crontab 
*/5 * * * * apache php /cacti/poller.php > /dev/null 2>&1
[root@b916fb282470 /]# exit
exit

Down/remove the containers & restore

kevin@p-serv-01:~/docker/cacti-test$ docker-compose down
Stopping cacti-test_cacti_1 ... done
Stopping cacti-test_db_1    ... done
Removing cacti-test_cacti_1 ... done
Removing cacti-test_db_1    ... done
Removing network cacti-test_default

kevin@p-serv-01:~/docker/cacti-test$ docker-compose up -d
Creating network "cacti-test_default" with the default driver
Creating cacti-test_db_1 ... done
Creating cacti-test_cacti_1 ... done

The /spine directory and /etc/crontab is empty.

kevin@p-serv-01:~/docker/cacti-test$ docker exec -it cacti-test_cacti_1 /bin/bash
[root@5a93e723b394 /]# ls /spine/
[root@5a93e723b394 /]# cat /etc/crontab 
SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root

# For details see man 4 crontabs

# Example of job definition:
# .---------------- minute (0 - 59)
# |  .------------- hour (0 - 23)
# |  |  .---------- day of month (1 - 31)
# |  |  |  .------- month (1 - 12) OR jan,feb,mar,apr ...
# |  |  |  |  .---- day of week (0 - 6) (Sunday=0 or 7) OR sun,mon,tue,wed,thu,fri,sat
# |  |  |  |  |
# *  *  *  *  * user-name  command to be executed

[root@5a93e723b394 /]# exit

Here is the docker-compose.yml file I used for this test:

kevin@p-serv-01:~/docker/cacti-test$ cat docker-compose.yml 
version: '2'
services:
  cacti:
    image: "smcline06/cacti"
    ports:
      - "8008:80"
     #- "443:443"
    environment:
      - DB_NAME=cacti_master
      - DB_USER=cactiuser
      - DB_PASS=cactipassword
      - DB_HOST=db
      - DB_PORT=3306
      - DB_ROOT_PASS=rootpassword
      - INITIALIZE_DB=1
      - TZ=America/Los_Angeles
    volumes:
      - cacti-data:/cacti
      - cacti-backups:/backups
    links:
      - db
  db:
    image: "percona:5.7.14"
    ports:
      - "3306:3306"
    command:
      - mysqld
      - --character-set-server=utf8mb4
      - --collation-server=utf8mb4_unicode_ci
      - --max_connections=200
      - --max_heap_table_size=128M
      - --max_allowed_packet=32M
      - --tmp_table_size=128M
      - --join_buffer_size=128M
      - --innodb_buffer_pool_size=1G
      - --innodb_doublewrite=OFF
      - --innodb_flush_log_at_timeout=3
      - --innodb_read_io_threads=32
      - --innodb_write_io_threads=16
    environment:
      - MYSQL_ROOT_PASSWORD=rootpassword
      - TZ=America/Los_Angeles
    volumes:
      - cacti-db:/var/lib/mysql
volumes:
  cacti-db:
  cacti-data:
  cacti-backups:

Thanks,
Kevin

[scline]

Hey Kevin, thanks for the report and that does not sound expected. Can you tell me what version of docker and docker-compose you are running?

It's as if the compose down command is deleting the container instead of just turning it off. If you still have the environment, after you docker-compose down can you paste the output of docker ps -a | grep cacti?

[kevburkett]

I think it's normal that docker-compose down stops and deletes the container (see reference here).

kevin@p-serv-01:~/docker/cacti$ docker-compose -v
docker-compose version 1.25.3, build unknown

Creating container:

kevin@p-serv-01:~/docker/cacti-test$ docker-compose up -d
Creating network "cacti-test_default" with the default driver
Creating volume "cacti-test_cacti-db" with default driver
Creating volume "cacti-test_cacti-data" with default driver
Creating volume "cacti-test_cacti-backups" with default driver
Creating cacti-test_db_1 ... done
Creating cacti-test_cacti_1 ... done

Running docker-compose down:

kevin@p-serv-01:~/docker/cacti-test$ docker-compose down
Stopping cacti-test_cacti_1 ... done
Stopping cacti-test_db_1    ... done
Removing cacti-test_cacti_1 ... done
Removing cacti-test_db_1    ... done
Removing network cacti-test_default

Containers fully removed:

kevin@p-serv-01:~/docker/cacti-test$ docker ps -a | grep cacti
kevin@p-serv-01:~/docker/cacti-test$

[scline]

Ah re-reading on those commands this is actually expected.

docker-compose down

Stops containers and removes containers, networks, volumes, and images created by up.

If your intent is to destroy the environment and on docker-compose up it will create a new one from scratch then docker-compose down -v should be used. This will remove the volumes and reinstall everything on the next up. You can continue to use DOWN and maintain SPINE/CRON files by adding the following volumes to the main cacti portion

From

    volumes:
      - cacti-data:/cacti
      - cacti-backups:/backups

To (have not tested yet)

    volumes:
      - cacti-cron:/etc/cron
      - cacti-spine:/spine
      - cacti-data:/cacti
      - cacti-backups:/backups

If the intent (I assume this is what you want) to simply turn off the containers and retain its information, then:
docker-compose stop

Stops running containers without removing them. They can be started again with docker-compose start.

is the command you would want to use.

[scline]

In the meantime, I can add some extra checks to the start.sh bootup script to validate SPINE and the cron is present. In the event, they are not re-compile and install w/o going through the whole "install cacti" loop

[kevburkett]

I actually tested earlier using these volumes and it works, but not sure if it's a good idea to be saving the entire /etc directory as a persistent volume.

    volumes:
      - cacti-data:/cacti
      - cacti-backups:/backups
      - cacti-spine:/spine
      - cacti-etc:/etc

Awesome, that would be great if the start.sh script can do that. It would save having to use these extra volumes. Thanks for the help with this!

[scline]

This has been updated in the most recent PR. Unfortunately, I did need to add a new volume for /spine, this was required in the event a user updates the instance utilizing ./upgrade.sh.

Statically assigning the built-in Spine version could cause version mismatches otherwise.