Running InfluxDB inside Docker container is extremely easy with the official image available on Docker Hub https://hub.docker.com/_/influxdb/. Things get a little more complicated when it comes to creating and restoring Dockerized database.
This was tested with InfluxDB 1.3 and Docker 17.11 on Ubuntu 16.04, of course it should work regardless of you OS or Docker version.
Creating InfluxDB backup from a running Docker container
This is a simple scenario in which we have a running Docker container with InfluxDB inside. Our task is to perform a backup of a selected database.
Get container ID
First of all you should get an ID from a running InfluxDB container. You may check it with a
docker ps command (CONTAINER ID is the first column):
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES e94f5db005a2 influx1.3 "/entrypoint.sh" 3 hours ago Up About an hour 8083/tcp, 0.0.0.0:8086->8086/tcp influx_1
Create a backup
It's easy, just execute:
docker exec -it e94f5 influxd backup -database mydb_to_backup dirname
- e94f5 - is your InfluxDb container ID
- mydbtobackup - is a name of database you want to backup
- dirname - is the name of internal container folder where DB backup will be placed (just make sure it's an empty directory, it will be used in the next command)
Then copy backup files to local directory with:
docker cp e94f5:dirname local_backup
- e94f5 - is your container ID
- dirname - is you container directory with backup files
- local_backup - is your folder path on the local machine
After executing this two commands you should have all backup files inside local_backup folder. Now it's time to restore this backup to the other InfluxDB container (which may be placed on the other host e.g. some VM on Azure or AWS).
Restore backup on Dockerized InfluxDB
Let's get to the restoring process which is a little more complicated than the backup. I assume that you already know how to get the ID of the target container with another InfluxDB. The overall recipe for the restoring is to create a temporary container which won't use Influx as an entrypoint (detailed process is described below).
First of all, you have to stop the container with InfluxDB. This is important to ensure correct database restoration process (it might be possible in future to restore database on a running container but it's not possible now):
docker stop e94f5
Of course, you should replace
e94f5 with your CONTAINER_ID first few characters.
Now it's time for magic. You have to create new Influx container which will not run Influx as a root process. Here is the command to do this:
docker run --rm \ --entrypoint /bin/bash \ -v influx_data_dir:/var/lib/influxdb \ -v local_backup:/backups \ influxdb:1.3 \ -c "influxd restore -metadir /var/lib/influxdb/meta -datadir /var/lib/influxdb/data -database [DB_NAME] /backups/[BACKUP_DIR_NAME]"
- influx_data_dir - is the directory which is used by your stopped Influx container to store DB info
- local_backup - is the directory with DB backup
- db_name - is name of the restored DB
- backup_dir_name - is name of database backup folder inside local_backup directory
Wait a few moments (or longer), some restoring messages should be visible in the output.
Then your database is restored you may start you stopped Influx container:
docker start e94f5
Everything should be fine. Your database should be available, which you may check via Influx CLI.
Some caveats encountered during the work with Docker backup/restore process.
Usage with Docker Volumes
In case of using Docker Volumes make sure that volume name specified for /var/lib/influxdb is correct. The easiest way to discover Volume name used by a running container is executing:
docker inspect e94f5 | grep volumes
- e94f5 - is your primary/running Influx container ID
Here is the example output:
Which means that your data volume name is my_influx-data.
Places where I've found informations which help me write this tutorial: