How to create a cold standby for confluence

Question

Hello,

I install confluence 4.3.5 on my Linux server with MYSQL as database. Now I would like to create a standby system. This is what I have done but I am getting a cluster error.

Make a backup of the production DB

Make a backup of the Production home dir

Install confluence on my standby server using the developer license

restore the production DB to the standby system

restore the production home dir to then standby system

Setup Rsync to copy the production home to standby home

Setup database replication from production to standby

restart confluence

This is the error I get::

Non Clustered Confluence: Database is being updated by another Confluence instance

Please advise

thanks

LuizA · Accepted Answer

Hello There,

Per the description that you posted I would suggest you to perform the following steps to create a test/backup environment of your instance:

https://confluence.atlassian.com/display/CONF43/Migrating+Confluence+Between+Servers#MigratingConfluenceBetweenServers-HowtoCreateaTestorDevelopmentSite

Then after you have your test/backup environment you can establish a kind of production backup strategy to perform the backup of your production instance, that would be a backup of the following:

* Backup of the Confluence's Database
* Backup of your Confluence-home Directory

By your comments it seems that you already have a kind of procedure for your backups, but anyway I would like to share with you an article about Confluence backup that would be interesting to take a look:

https://confluence.atlassian.com/display/DOC/Production+Backup+Strategy

So based on this, if you have the backup of the above items you'll be able to restore your Confluence into your cold standby instance. The point that you're receiving error is probably due to the database replication, that is a kind of procedure not recommended for the Confluence database, normally most of our customers just establish a backup routine and then if needed they restore the backup from a specific time stamp and not using replication from the production database. And about the Rsync this is not the problem here as it is just a copy procedure from the file system, not the case here, but our main problem is the replication of the database.

In this case I would suggest you to create a backup routine of your Confluence-home directory (file system) and Confluence's database, and then restore from a specific point if needed in the future and do not use this kind of replication from the database side to avoid this kind of error, this is what we usually suggest to our customers.

Hope this helps.

Kind Regards,

LJ.