Information Missing from Cluster Monitoring

[Darren Shinkins]

Hi All,

I'm currently testing an instance of Jira data centre (v8.6.1) with 6 application nodes in a cluster. It's all set up and working correctly, with the exception of Cluster Monitoring

Steps to Recreate

* Navigate to system "Clustering" admin page

* Enter password gain access to temporary admin privileges

Expected Result

Page loads and lists details of all active nodes in the cluster, including:

* Node ID

* Cluster Address

* Load

* memory

* Uptime

Actual Result

Page loads and lists active nodes in the cluster, but the following details are missing:

* Load

* Memory

* Uptime

Actions Taken So Far

Implemented workaround to bug with Hipchat classes 

The workaround described in bug ref JRASERVER-70711 suggests deleting records from the pluginstate table. That has helped somewhere; prior to following this instruction, the page was blank.

There is one more similar entry in this table:

com.atlassian.hipchat.plugins.core 

Not sure if this should be deleted too?

Does anyone have any other suggestions on what may be causing the problem at all?

Many thanks

Darren

[KAGITHALA BABU ANVESH]

Hello @Steve Mills ,

Recently we faced the same issue. 

We cleared old logs and some unused files and old backups to get free space. When the free space is greater that 50% the error was cleared. All nodes showing information in clustering page.

It is a workaround, Not sure the exact solution.

Thanks,

Anvesh Kagithala

[Steve Mills]

Did you ever find a solution to this?  We began experiencing the same thing about 2 weeks ago.

We do not have the Hipchat entry in the pluginstate table as mentioned.

[Darren Shinkins]

Hi Steve,

For us, deleting entries from the pluginstate table partially fixed the problem. In Jira Software 8.6.1 showed the entries for the nodes on the Clustering page, but no further details (load, memory usage etc).

This is apparently a known issue with that version of Jira, and upgrading to Jira Software 8.13 fixed the issue for us.

Hope that helps!

[Steve Mills]

Thank you for your prompt response.  We checked our pluginstate table in our lower environments and production (we are only having the issue in production) and they are consistent.

Our free disk is in the installation directories is between 58 and 61% free across our 3 nodes.

I appreciate all suggestions.

We do have an open support case with atlassian and have tried a couple things, but it is still flakey after a restart.  But after 30 - 90 minutes, all nodes show up and clustering appear to work from there.