![]() |
![]() HP OpenVMS Systemsask the wizard |
![]() |
The Question is: We recently had a hang of our VMS 5.5-2 system after a significant loss of power event on some of our network devices thru which we communicate via DECnet. We lost power to 5 terminal servers and 2, 8-port FDDI hubs. During the power loss there was no d isruption of the VAX 4500 nor it's disks. My question is in regards to DECnet's response to loss of network devices. How quickly does DECnet timeout theses devices? Can significant power distruptions to network partners cause the VMS hang that we obser ved? The Answer is : Without detailed specifics of the power outage problem and of the hang, it is very difficult to determine exactly what happened. DECnet is not used to communicate with terminal servers -- the MOP protocol (available as part of DECnet Phase IV, DECnet-Plus, and in the OpenVMS LANCP utility available in OpenVMS V6.2 and later) is used for the initial download and (if necessary) for an upline dump. Other than these MOP operations that occur at terminal server bootstrap and at terminal server crash, LAT or (depending on the terminal server) IP protocols are used. Not DECnet. Failures at FDDI hubs can obviously cause problems for any network protocols operating over the FDDI link. This includes DECnet, LAT, IP, clustering, and any of the other protocols that might be using the FDDI rings. DECnet circuit timeouts should detect and report circuit failures, please see the current system NCP or NCL settings for the setting of the timeout on the local system. The OpenVMS Wizard would not typically expect a complete hang due strictly to DECnet.
|