Server Administration

This chapter describes the day-to-day administration tasks that can be performed by the operator of the virtual machine. The application configuration, the management of the service accounts and the application security updates remain the responsibility of beCPG.

Start and stop the beCPG service

Use the commands:

cd /opt/becpg-srv-instances/inst1
docker compose stop
docker compose up -d

or:

cd /opt/becpg-srv-instances/inst1
docker compose restart

If the optional OLAP or AI modules are installed, prefer ./tools/instance.sh restart, which takes every deployed component into account.

Check the state of the services

cd /opt/becpg-srv-instances/inst1
docker compose ps

The STATUS column must show healthy for the components declaring a health check. A component that stays in error must be reported to the support team.

Server logs

Use the commands:

cd /opt/becpg-srv-instances/inst1
./tools/instance.sh logs

or:

cd /opt/becpg-srv-instances/inst1
docker compose logs -f -t --tail="100"

or view the logs of a single container, here the becpg-mail container:

docker compose logs becpg-mail

To learn more about logs, go to the official documentation.

Additional logs are available in each container in /var/lib/docker/volumes/*_logs/.

Show the memory usage

cd /opt/becpg-srv-instances/inst1
./tools/instance.sh stats

Diagnose the instance

In case of slowness or abnormal behaviour, a full diagnostic (containers, memory, disk, index) can be produced and sent to the support team:

cd /opt/becpg-srv-instances/inst1
./tools/instance.sh diagnose

Show available space on the system

df -h
du -ms /var/lib/docker/volumes

Increase the data partition

Increase the disk in your virtual machine hypervisor.

Then use the following commands (fdisk -l lists the partitions):

sudo reboot
# Replace sdb with the partition to resize
sudo resize2fs /dev/sdb

Fix indexing in case of a crash

cd /opt/becpg-srv-instances/inst1
./tools/solr.sh fix

Reindex in case of server crash and index corruption

Perform reindexing if the fix does not correct the problem.

cd /opt/becpg-srv-instances/inst1
./tools/solr.sh reindex

Warning: reindexing can take a long time (more than 24 hours)!

Administration