Offline access to some Confluence content during maintenance

[Thomas Ehardt]

Adoption of Confluence in our organization has grown to include runbooks for our operations team; these runbooks contain information needed to handle various situations.

While Confluence is down for maintenance, these runbooks are not available to the operators and therefore they are in the dark as to how to handle issues. I'd like it to be that we can take Confluence down for maintenance and still have the information available for viewing.

We have Apache in front of Confluence, so my first thought is add a caching layer and prime the cache with the appropriate pages (it's a small subset of all our content). This way, when Confluence is down for maintenance, the operators can still access the information they need when they need it. It is only a subset of documents that need this functionality.

Has anyone else run into a similar situation? If so, how did you solve it?

[Jason Hensler · Accepted solution]

I have two ideas for you.

The first option would be to setup a "failover" confluence similar to this jira guide >https://confluence.atlassian.com/display/ATLAS/Failover+for+JIRA

You would want to setup an additional confluence server and have it contain the "Ops" space. Then you would have apache mod_proxy or mod_jk "loadbalance" the production server with the "Ops" server. You can setup the mod_jk (not sure about mod_proxy) to only send traffic to the "Ops" server when the production server is down so, that traffic would be routed to the "Ops" server without configuration changes. I believe that based on the guide, you could do would be able to stay with-in licensed terms if you stopped your dev server, brought up the "Ops" server, stopped production and made changes, brought the production server back up, stopped the "Ops" server, and restarted the normal dev server. or you could stop production, bring up the "Ops" server, patch production, Stop the "Ops" server and bring Production back up but, this would leave some unavlibility of the Ops documents

The second would be to export the space as html via a script or the spaces "Avanced" menu. Place this on the apache server and activate a mod_rewrite (or someother redirection, could be a custom 503 error page the reads the request url) option to redirect the request to the html pages.

[Volker_Kleinschmidt]

You can export specific pages to PDF or Word format, and you could even automate that for regular updates. Seems for the few documents the operators need this would be the simplest solution.

[Thomas Ehardt]

Originally, we had Word documents out on a shared LAN drive. The ease of use with Confluence and the fact that we could standardize documentation (through Scaffolding templates) led us to push very hard for getting the documentation into Confluence.

However, due to bureaucratic silliness, changing the ops environment documentation to point to PDF or some other type of file is considered a production change, and we're talking documentation for roughly 1,000 - 1,500 jobs.

We're also fighting the battle of SharePoint vs. Confluence and anything we can do to get Confluence adoption spread is a good thing.