HA configuration for Bit Bucket Data Centre

[John Lawlor]

Looking at  this article:

https://confluence.atlassian.com/enterprise/bitbucket-data-center-668468332.htm

There seem to be several single points of failure in this setup. The database, the NFS mount and the proxy server. What is supposed to happen should one of those go down?

[Christian Glockner]

Hi John,

Data Center achieves high availability for Atlassian applications such as Bitbucket. You are correct that - unless other provisions are in place - the shared storage, the database server, and the load balancer become potential single points of failure in such a setup.

We leave it up to our customers to decide whether and how they want to ensure that these pieces of infrastructure are equally highly available, but it indeed really only makes sense to use Data Center if they are resilient too.

If the load balancer goes down, Bitbucket Server will be unavailable as no requests from users will be routed to it. If the database goes down, Bitbucket Server will be unavailable because virtually all operations in Bitbucket Server (including logging in) require the database to be available. And the same applies for storage - if that goes down, then at the very least all SCM operations (i.e. git operations) will fail, but this may also affect parts of the UI, such as viewing sources, diffs, pull requests, etc that touch on source code.

Cheers,

Christian Glockner

Premier Support Engineer

Atlassian