0

Can anyone provide me with an alternative along with the steps to install Redmine HA instances?

I tried to solve it with Docker swarm, but it seems very sophisticated.

Aravindh
  • 13
  • 5
Alexey Gorbachev
  • 25
  • 3
  • 1
    Tutorial recommendation questions are off-topic for StackOverflow. Also, not sure what you mean by "alternative steps" - you haven't shown anything regarding the actual steps you're taking. (plus, once you rework this question to contain more details, it would fit much better on ServerFault). – David Makogon Nov 17 '16 at 12:57
  • "alternative steps" - means Kubernetes or something. Is it possible to install Redmine in Docker swarm without discovery service? – Alexey Gorbachev Nov 17 '16 at 13:36

1 Answers1

5

The Redmine application itself is (almost, see below) stateless since all state is kept in the underlying database. You can easily run several instances of Redmine on different servers/VMs, all pointing to a single database (cluster) by specifying the database in database.yml. The only remaining part would be the file attachments stored in files/ within Redmine. You would have to put this on a network share.

Any loadbalancing capable reverse proxy (e.g. HAProxy) can then be used to distribute requests across your Redmine instances or to implement a failover in case one of them goes down.

To make your database highly available, you'd have to look into the replication capabilities of your chosen DBMS. For instance, PostgreSQL has a section about HA in its docs.

It's not black magic but it requires continuous work to make Redmine run fast and reliable at large scale. My company Planio Redmine Hosting does this day in and day out.

jan
  • 1,130
  • 1
  • 9
  • 11