JIRA DC Node: ehcache connection refused

[Sebastian Kouba]

I'm trying to start a second JIRA DC node but I get the following error on the second node (jira1):

[n.sf.ehcache.CacheManager] Cache com.atlassian.jira.crowd.embedded.ofbiz.EagerOfBizGroupCache.groupCacherequested bootstrap but a CacheException occured. Error bootstrapping from remote peer. Message was: Connection refused to host: 192.168.100.18; nested exception is:
 java.net.ConnectException: Connection timed out (Connection timed out)

and a bunch more connection refused / timeout messages.

According to my telnet test, port 40001 is open between the servers. This was run on the second node, jira1, where the error occurs.

telnet jira0 40001
Trying 192.168.100.18...
Connected to jira0.

I've only set the node names in cluster.properties so everything else is default.

What other ports need to be open for DC to work? There is a mention in the documentation of "unrestricted access" between machines but I'm not sure that's going to happen in my case...

I've tried deleteing the plugin caches as per help article, no dice. The first node is running as intended.

Any ideas?

Thanks

=> Solution

There is another port other than 40001 that needs to be opened. You can add this line to cluster.properties:

ehcache.object.port = 40011

[Sebastian Kouba · Accepted solution]

We've temporarily disabled all firewalls on the network. Now the nodes work.

Can anybody provide information as to which ports need to be acessible between machines? Looks like some ports in the 30000-50000 range.

Just found this in the documentation:

"Nodes must be configured to have unrestricted port access to each other"

Really? Just open the flood gates?

[Sebastian Kouba]

Seems related to this after all:

https://jira.atlassian.com/browse/JRASERVER-64974

[Sebastian Kouba]

Very disappointing that for this "enterprise" product there is no documentation of some of the most basic setup steps, except hidden in some ticket to update the documentation...

Quote from the ticket linked above:

New variable: ehcache.object.port
Default port: 40011

Note that 40011 doesn't actually seem to be the default port. I had to set the port in cluster.properties before it worked as expected.