KBEA-00019 - Fixing Agents that do not reregister with the Cluster Manager

Summary

The agent is upgraded or installed and appears to run. However, the Cluster Manager does not list the agent.

Solution

Look at the agent logs (/var/log/ecagent*.log on Solaris, /var/log/ecagent*.log on Linux, c:\ECloud\ecagent*.log on Windows) and check at the bottom for any connection errors. Common errors are:

• Unable to resolve the name through DNS. Enter the IP address for the Cluster Manager instead using ecconfig.
• The Cluster Manager name is mistyped. Fix the Cluster Manager name using ecconfig.
• The Agent or the Cluster Manager has a firewall that limits access to port 80 and other ports required by ElectricAccelerator. Set the firewall to allow the agent process to access any port.

Examples

Running ecconfig as root/Administrator:

> ecconfig
Cluster Manager     = lonestar-cm:8030
Number of Agents    = 2
Is LSF Host         = n
Secure Console Port = n
Agent Temp Dir      = /tmp
> ecconfig \-cm mycm:8030

Applies to

• Product versions: 4.x and later
• OS versions: All