Thanks for your votes and comments on this issue.

After discussing this internally, we've revised our policy on virtualised hardware for Confluence. Since we don't stipulate which physical hardware you can run Confluence on, we have decided to apply the same approach for virtual hardware.

Atlassian will support installations of Confluence on a virtualised environment (including clustered installations), but can only assist with issues that aren't related to or caused by the virtual platform on which it is running.

Rather than recommending specific platforms for running Confluence in a VM, we recommend sticking with a technology that you are familiar with and can manage well. Our policy page has been revised to list some general recommendations that you should be aware of when running Confluence in such an environment.

+1 on the above comment.

If Confluence wiki is positioned as an enterprise wiki product, clustered in a virtualized environment support is a MUST. The is a critical requirement and we need to have a timeline on this. can we have an update? Thanks.

We have a critical project deadline coming up and have been pushing for Confluence as our preferred solution, in part on it's ease of install, configuration, feature set, and use. But if we need to install this on physical hardware then we're going to be six (6) to eight (8) weeks behind before we even get started! It takes time for the hardware request to get through the appropriations process and requires VP approval. We were really counting on our ability to very quickly stand up a cluster of Confluence front-ends as virtual machines and spend our time customizing it and training the users on how to leverage the expansive feature set.

The "we don't support VM based cluster" stance may well be a show-stopper here; forcing us to kluge together a solution from currently implemented applications. It won't be nearly as good and definitely not what we were hoping to present to the end-users ... but we just don't have the time and resources for a hardware based cluster.

Any update on the status of this issue? Support of clustered Confluence on VMs is critical to us at Boston College. All new server requests have to be VMs, unless there is an overwhelming reason that it must be on bare metal. To me it seems like a no brainer for you guys to create Virtual Appliances based on Confluence. Then end admins download a fully functional VM that they can slap into Citrix XenServer or VMware. Then all the admin need to do is start the VM, rename and configure for their network, start confulence, add a license, point it at a database and Confulence is up an running

At Thomson Reuters, We need support for Confluence clustered in a virtualized environment, else soon we will not be able to support our expanding usage of the wiki.
Why would confluence not support this? Virtualization is certainly a trend you can't ignore for long.
I would also like to see support for web acceleration. caching for web accelerator adversely effects the editor.
thanks
-steve dungy

We received this comment from a customer on how he got the cluster to stop throwing cluster panics on VMWare:

I digged into the Oracle Coherence documentation and forums. A few things I changed in the VM environments to improve the multicast were extra NICs for the dedicated multicasting on a separate switch with multicast enabled and increased switch buffer.

From Atlassian's perspective the environment is still unsupported.

At this stage, a feature like this is not in planned in our short or long term roadmaps. We will continue to observe feedback from our customers on this issue and update this feature request if any progress is made.