How To Control Single Server In Distress Causes Cascade Of Members To Leave From Healthy Physical Servers or VMs? (Doc ID 2623010.1)

Last updated on AUGUST 28, 2023

Applies to:

Goal

For an example when experienced a hardware issue with a single physical server/VM in the Cluster. The nature of the hardware failure caused the physical server/VM to be intermittently reachable by the Members hosted on other physical servers/VMs participating in the Cluster. The usual user expectation is that a failure on a single physical server/VM should NOT influence the coherence members of other physical servers/VMs. However, out-of-the-box coherence cluster configuration appears that the failure of members a physical server/VM to talk to other cluster members convinced the senior member that those (healthy) members had left the Cluster. Then when the senior member heartbeat did not contain those healthy members, they do in fact actually leave the Cluster.

How To Control Single Server In Distress Causes Cascade Of Members To Leave From Healthy Physical Servers or VMs? 

Other relevant questions to this are...

• How can a Split-Brain event be detected?  
• How are they defined?
• How does Oracle Coherence deal with Split-Brain?

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.