Atlassian Bamboo Failover recommendations

[Chris Riley]

I cannot find any documentation about architecting a Atlassian Bamboo configuration for failover. The only failover documentation I have found relates to JIRA. Can you provide any suggestions?

Based on what I have read and seen with other clients is the creation of a cold standby who's file system is synced with the primary bamboo instance. What isn't clear is whether we can have a clustered DB that both the primary and cold standby use or if they require there own independent DB.

Thanks

Chris

[Felipe Kraemer · Accepted solution]

Hi Christopher,

What isn't clear is whether we can have a clustered DB that both the primary and cold standby use or if they require there own independent DB.

According to the JIRA Failover documentation, all databases that are supported by JIRA feature replication functionality that allows establishing master/slave replication or even clustering. Databases like PostgreSQL, Microsoft SQL Server and Oracle support synchronous replication - some even with a failover mechanism of their own.

Bamboo supports these databases as well, so it should be okay to apply the same suggestion given in that documentation for Bamboo.

I also found this post which you may find useful.

[Chris Riley]

Thanks for the clarification on the DB portion. The client is considering MS SQL or Oracle. I have told them that Oracle RAC would be preferred option (failover or cluster mode). I am not as familiar with MS SQL but assume a cluster/failover solution is similarly defined. We will be using a JDBC URL that defines the cluster nodes in case of DB failure. Separate from the DB what I am gathering here is that Bamboo will operate similarly to JIRA and the JIRA best practices document is the guide that can be applied both JIRA and Bamboo?

[Harry Chan]

It's similar, but Bamboo runs embedded Jetty and not Tomcat, so there are differences. Cold standby still stands though.