jump to navigation

Oracle Coherence Enterprise Management April 19, 2009

Posted by shawngandhi in Technology.
Tags: , , ,

Last week I attended the Oracle Coherence Special Interest Group in NYC. They introduced a much needed tool that I’ve been waiting to check out since I first started toying around with Coherence: the Oracle Management Pack for Oracle Coherence.

I finally got a chance to play around and see it in action. When using Enterprise Manager 10g, you now have the ability to deploy, manage, and maintain multiple coherence clusters all from a simple web interface. It works by connecting to coherence clusters using JMX.

The management dashboard allows an administrator to monitor the health of your cluster from a single point. You can compare the performance of various nodes, caches, clusters, and connections. Upon the detection of a faulty node, you can simply evict it from the cluster so it can’t cause any damage.

I was happy to find the ability to monitor publisher/receiver success rates, which are conveniently graphed to show outage severity throughout the day.

All cluster metrics are stored in an Oracle database (which comes included with the management pack) and can be queried for customized cluster statistics. The manager gives you the ability to configure alerts and actions based on data metrics gathered by the system. I spoke with one of the lead developers, Debu Panda, who said they are very open to displaying additional data that customers feel would help monitor and manage their cluster.

The Coherence Management Pack is more than just a health monitoring tool. They are working on a feature which will allow you to update jars/binaries in your cluster by uploading them from the dashboard and perform systematic upgrades of nodes in the cluster in a rolling restart fashion. You can even create and deploy new nodes if the cluster is reaching certain thresholds. Unfortunately, this will not adjust the segmentation count automagically — that’s a feature planned for Coherence 3.5.

I wasn’t pleased to find that the finest possible granularity of data collection is a minute. Although I can see that being fine for some customers, I can envision several real-time projects that would find this to be a constraint.

Secondly, they do not have the ability to view all the data in the cache. This would be helpful to analyze how the cluster is being used from a data perspective. However, I can understand that such a feature comes with its own set of problems and Heisenberg effects.

Overall, its great to finally see a tool that can manage the cluster as a single system. Runtime configuration management, historical data, customizable metrics and alerts, all combined into one simple product really make the task of provisioning and managing your Coherence clusters less daunting. I definitely give this tool a thumbs up.



1. Craig Blitz - April 20, 2009

Thank you for the post and the feedback Shawn. Look forward to your continuing involvement with the New York Coherence SIG.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: