This Repository / Docker-Image makes it easy to start hosting a Atlassian Confluence - Instance. It is located at the following places:
- Docker-Image: https://registry.hub.docker.com/u/derjudge/confluence
- Git-Repository: https://bitbucket.org/Judge82/docker_confluence.git
I started this project as a fork of cptactionhank/atlassian-confluence, because I wanted to change some things in it:
- cptactionhank's version does not contain a DB like PostgreSQL or MariaDB/MySQL. This is not a lack! Actually, it is good style and recommendable, since Docker's concept is to link a DB-Container if you need one. But I wanted it as an as-simple-as-possible way to get started with Confluence, without having to care about setting up and linking another DB-Container.
- The oldest version of Confluence available in cptactionhank's image is 5.5. The reason I started this image is that I had to migrate an existing Instance of Confluence which was powered by version 3.5.13 of the software to the recent version (which is 5.7.4, at the time of this writing). That's why all the ancient tags/branches are contained in this container.
- As of October, 20th 2015 this container doesn't use the java parent image anymore but debian stable.
There are multiple branches named like the corresponding version of Atlassian Confluence it contains. Please see Docker image's tags to get a list of all versions supported so far.
This image offers a tag latest
and several tags named in a X.Y.Z
schema.
The later are meant to be used when needed in special scenarios like migrations to a more recent version, starting from these older ones, restoring backups from that specific version or staying on that version due to an expired license; they are not intended to be used longer than temporarly in production environments! Thus, their states are kind of freezed in the state they are when initially created to keep backward compatibility as best as possible. No updates or enhancements, neither security nor features, will be backported from latest.
This is why latest
is the only recommended tag to be used in production environments.
Since this image is for the impatient fellows out there, let's start with the fastest and easiest way to bring up the most recent version of Atlassian Confluence to life. Just type the following on any Docker-Host:
docker run -p 8080:8090 -d derjudge/confluence
Notice: This is true for all versions, instead of 3.5.*
. Use the following
for these instead (replace 3.5.17
with the version you want to start; mind the changed container port):
docker run -p 8080:8080 -d derjudge/confluence:3.5.17
Both variants will take some time to load. You might notice an increased use on
system resources; especially CPU usage will temporary grow up to 100% on one or
more cores. The process generating that load is java
. This is normal and
perfectly OK, because this means Confluence is starting up and initializing
itself. Based upon the performance of your Docker host, this might take several
minutes to complete; on moderate hardware, usually not longer than 5 minutes.
If it takes noticeable longer, you might want to consider using newer hardware
to host Confluence.
You can determine that everything has started by the following:
- CPU usage lowers significantly
- the line
INFO: Server startup in X ms
appears in/usr/local/atlassian/confluence/logs/catalina.*.log
(inside the container; usedocker-enter/nsenter
to enter the container). - A ... "positive" line appears in
/usr/local/atlassian/confluence-data/logs/atlassian-confluence.log
. This will be something likeinit Plugin system started
orinit Confluence is ready to serve
. What you see there exactly heavily depends on the Confluence version you are starting. Again, I'm speaking about the file inside the container; usedocker-enter/nsenter
to enter the container.
Once everything has started, you should be able to point your browser to
http://<ip-of-your-docker-host>:8080
and see Confluence's landing page.
From this on, just follow Atlassian's official setup guide.
You can either use Confluence's internal, file-based database to have Confluence initiate it's data structure or use any other Database System, the container can connect to. This can either be a PostgreSQL or MySQL/MariaDB database you are hosting in your network or you can use the PostgreSQL database included in this image.
I (and Atlassian) strongly recommend to use PostgreSQL database or any external database, since the embedded database is neither very reliable, nor scalable. Also, performance might be quite bad, compared to a mature (and tuned) DBMS. The PostgreSQL database which is included in this image is not optimized for speed, but for reliability and compatibility with a low amount of available resources to support as much hosts as possible, even when they consist of older hardware or are quite limited in available resources. If performance is of major importance for you, please consider to use an optimized external DBMS instead.
The "Docker way" to do something like this would be to start another container running the DBMS and link the Confluence- and DBMS-Container. So it is considered "bad style" to have a DBMS included in this container from a Docker perspective. But since this image aims to get you started as fast and easy as possible, a PostgreSQL server was included on purpose, to not bother you with this additional step.
You can freely decide not to make use of it and go the "Docker way". I consider you are experienced enough with Docker to figure out how this is done, if this is what you want. So it won't be described here. You will find tons of How-Tos on the Internet.
However, if you decide to use the internal PostgreSQL database, choose and enter the following options and values in the corresponding dialogues of the Confluence setup steps:
Choose a Database Configuration
: ChooseExternal Database (PostgreSQL)
Configure Database
: ChooseDirect JDBC Connection
Setup Evaluation Database
: Enterpostgres
intoUser Name:
-Field. Leave all other fields "as is" (empty password).
If you have a backup of another Confluence instance, this easily grows up to several gigabytes in size; you have to count
all images, attachments and so on which grew into your Confluence datastore. Since this is usually too big to upload via
browser. Thus, the best way is to have Docker mount a folder of your Docker host filesystem, which contains this
backup-file, to the path /usr/local/atlassian/confluence-data/restore
of your Confluence container. This way, you do
not have to upload the file, but simply choose it from a list of available backups during Confluence setup steps.
To do so, simply attach the option -v /path/to/backup/file/on/your/host:/usr/local/atlassian/confluence-data/restore:ro
to the docker run
- command. The following is an example of a complete Docker run - command:
docker run -p 8080:8080 -v /home/mr/conf-bak:/usr/local/atlassian/confluence-data/restore:ro -d derjudge/confluence:3.5.17
That's all!