Helen,

Can we have have any visibility into why Atlasssian has decided to not explore this bit of functionality? It is too technically challenging? Do you not have the resources you require to write a high availability application?

The document you point to from one of your certified expert partners says it it possible and details the steps to approximate a high availability version of Crowd. It seems strange that out of one side of your mouth you are pointing to a solution from an organization under the Atlasssian umbrella and with the other you are saying that it is not a priority.

It is great that you are looking to provide tighter integrations with your other products, but this is of little solace when your single monolithic Crowd instance tanks and now your 20,000 users can't log in. This message makes it seem that Atlassian is glibly accepting of the fact they are building software with single points of failure. Not only that there is no visibility into WHY this decision is being made. High availably applications are hardly a new technology. Why is Atlassian purposely deciding to build non scalable software?

Hi everyone,
Thank you for your votes and comments over the years on this documentation request.

Ideally we would like to be able to meet everyone's requirements and needs for Crowd, but the reality is that this is not always possible. Unfortunately, this is one of those cases, particularly since we don't provide support for Crowd clustering.

Closing issues that have a fair share of votes and comments is not an easy decision to make. However given that this has been open for several years now and we don't have any foreseeable plans to implement this, we've decided to make this clear so you know sooner rather than later.

If you'd like to read more information and discussion of Crowd clustering, have a look at https://jira.atlassian.com/browse/CWD-1053

If you have any questions, feel free to leave a comment or contact me on helen@atlassian.com

Cheers,
Helen Hung
Atlassian Product Management

Hi Caspar,

We're running around in circles here. I found all of these links before plus some slides. And that's it.

There is no documentation currently online (read: indexed by google.com or archive.org), besides what's in CWD-2925, on the actual tweaks needed, e.g., to enable Tomcat Clustering and actual session replication (using the <distributable /> tag in web.xml). And possibly additional tweaks necessary for Hibernate. I've actually managed to get session replication to work now, and according to our tests it seems to work with and without sticky routes on the reverse-proxies, so that is good. However, what I would like is to check with the old "documentation" there was, whether I've missed anything.

Your first two links give no configuration details, just a high level overview. That's for management. I'm a system administrator and already got beyond the point where I'm drawing pretty diagrams. I'm testing it in the field.

Regarding Crowd load-balancing and Load balancing Crowd on Atlassian Answers are the only threads on that subject and they both link back to the original forum post as:

• 'Actual' current documentation: http://forums.atlassian.com/thread.jspa?forumID=104&threadID=25359

For the benefit of everyone else, I'm going to post an update to CWD-2925 with what we have so far and maybe to kick off the discussion again.

ustuehler1, I'm not sure what was available on those forum links, but you might find the following documentation from one of our partners helpful:

• Atlassian Crowd: High Availability For Demanding Environments
• Atlassian Crowd: A Simple Clustering Solution
• Atlassian Answers related to Crowd Clustering (and variations on such searches), which have replaced our forums; often you'll get direct answers from us Atlassians or our partners, but some answers are from customers who are relating their experiences and solutions.

I should point out that while we don't officially support these kind of configurations (yet - if you would like this to happen, please vote on CWD-1053 if you haven't already), we do know that some customers are using these kind of approaches anyway.

The forum thread is gone, now that the forums are gone, and extranet.atlassian.com is accessible only to Atlassian employees. Is there any archive of the forum threads, or another place in the public where the information is still accessible?

Hi everyone,

Just letting you know that we've updated the current Atlassian status in the description of this issue, and you should go check it out!

Eugene Mak
Atlassian Product Management

Here's a forum thread with some useful information:
http://forums.atlassian.com/thread.jspa?forumID=104&threadID=25359