| Title: | MAILBUS 400 User Forum |
| Notice: | kits 100-109 - Infocenter //www.digital.com/info/messaging |
| Moderator: | IOSG::MARSHALL |
| Created: | Thu Jun 11 1992 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 3210 |
| Total number of notes: | 9174 |
Hi all, I have a question about TCP/IP Support in a VMS-Cluster. We have a situation where a user agent (ALL-IN-1 V 3.2) runs on all CPUs of a homogeneous alpha-cluster (2 CPUs). The agent uses XAPI and XDS. The DSA and MTA run on one CPU only at a time. If this CPU fails, DSA and MTA will be started on the second CPU. The transport is RFC1006. We use DECnet alias and Internet alias (via UCX). Is it possible that connect-requests from the agent and from peer MTAs to the DSAs and MTAs presentation address is directed to the "wrong" CPU, i.e. the CPU where the DSA and MTA are NOT running ? Is it possible to force the connect-requests to the CPU where DSA and MTA are running and switch (transparently) to the second CPU if the first one fails ? regards, Erich
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 3198.1 | IOSG::MARSHALL | Fri May 23 1997 17:49 | 10 | ||
I know nothing about internet aliases so can't answer that part of your question. But if you use MTA sets, you should't need aliases. Define the two MTAs as an MTA set. Then other MTAs will connect to whichever of them is running. Note you still need manual failover to the second MTA. Also you can configure the second MTA to pick up messages in the failed MTA's queue, but I don't remember the details. Scott | |||||
| 3198.2 | yes, but... | OSITEL::rtont2.rto.dec.com::Erich | SI-Office | Mon May 26 1997 09:05 | 8 |
Hi, thanks for the answer. I would like to use a MTA set as the last chance because the set does nothing for me from the agent perspective, and I still need a solution for the DSA. regrards, Erich | |||||
| 3198.3 | FORTY2::PALKA | Thu May 29 1997 11:32 | 7 | ||
I think you will have to declare the DSA to have a presentation address
with multiple nsaps, one for each node in the cluster.
I believe the UCX alias IP address will not find the node in the
cluster which is running the DSA.
Andrew
| |||||