A specific Task Server was allowing computers to connect to it, but after one hour or less the clients (including the Task Server its self) disconnected and could not reconnect even though they tried. The only way to fix the problem was to restart the "Altiris Object Host Service".
After restarting the service the computers could reconnect, but after a period of time they disconnected again. Rinse, repeat.
The client logs showed the following series (excerpt) similar to the following:
Source: Client Task Agent Description: Could not register using "http:/SS-Main1.epm.local:80/Altiris/ClientTaskServer/Register.aspx"
Source: Client Task Agent Description: An attempt to register on Task Server [PCI3-SMPSS02.pci1-iroot.local] over [http] completed with status [FAILED (may retry)].
Source: Client Task Agent Description: Task Server operation failed: Web application could not reach Task Server Error code: 0x00000002, Error description: The handler 'Register' is failed to process request. Could not contact Task Server's service., Retry: Yes, Registration is lost: Yes
Symantec Management Plateform 8.5 RU2
The Task Server was a virtual machine. Something was wrong with the virtual NIC, or the NIC in the hosting machine being used as the root NIC configuration.
Instead of using a VM the administrator brought up a physical box, as a test, to replace the problematic Task Server. After configuring it as a Task Server the problem never reoccurred again.
He opted not to identify the root cause in the VM, since the new physical box worked so well; but I is probably that the deleting the virtual NIC in that VM and creating it again, would have helped the situation. Wireshark showed a lot of dropped packets coming out of the NIC during the diagnostic phase of troubleshooting.
Subscribing will provide email updates when this Article is updated. Login is required.