Operating System - OpenVMS
cancel
Showing results for 
Search instead for 
Did you mean: 

tcpip route could not be fixed

Gesmundo
Advisor

tcpip route could not be fixed

We have this cluster of blade servers running OpenVMS 8.3-1H1 on 2 sites. Everytime we change the default route on a server on site A, a server on site B which boots later will lose its default route. We can run the TCPIP$CONFIG on the server to change its default route but the server on site A that is booted will lose its default route. We have a duplicate of this cluster and we can change the defaut route from any server and other servers can boot without losing its default route. We can see that the servers share the same TCPIP databases (SYS$COMMON:[SYSEXE]TCPIP*.DAT). We tried create a new route database (SYS$COMMON:[SYSEXE]TCPIP$ROUTE.DAT) but the issue persisted. How can we fix the default route on all the servers? Thanks. Noel
3 REPLIES
Hoff
Honored Contributor

Re: tcpip route could not be fixed

OK, so you have a cluster or two running comparatively old OpenVMS and TCP/IP software, one of which is not maintaining the default route across servers within the cluster, and another cluster which apparently is maintaining the routing.

 

What TCP/IP Services version and patch levels are in use for the various servers involved here?   (You do mention duplicates, but are those hosts in the second cluster manually created copies, or are they copies imaged from the problematic configuration?  Identical hardware for both clusters, or are there different models or types of Itanium servers in use?)

 

What command(s) are you using to change the default route on the working and not-working configurations?

 

What command(s) are you using to show that the default route has been lost?

 

What are the differences between the working and the not-working cluster configurations?

 

Are BGP or other routing protocols in use?

 

Any reason that you're not running OpenVMS V8.4 with TCP/IP V5.7; something a little more recent than V8.3-1H1?

 

I'd probably look to use the blade inboard NAT — HP has a fancy marketeering name for this NAT — to avoid having to reset the network, too.

 

 

Gesmundo
Advisor

Re: tcpip route could not be fixed

Hoff, Thanks. >>What TCP/IP Services version and patch levels are in use for the various servers involved here? (You do mention duplicates, but are those hosts in the second cluster manually created copies, or are they copies imaged from the problematic configuration? Identical hardware for both clusters, or are there different models or types of Itanium servers in use?) The system is using TCPIP Services V5.6 - ECO 5. The 2 cluster have exactly the same hardware except for the storage systems. The erring cluster is an image copy of the running cluster. >>What command(s) are you using to change the default route on the working and not-working configurations? We run the SYS$STARTUP:TCPIP$CONFIG to change the routing. >>What command(s) are you using to show that the default route has been lost? $ UCX SHO ROUTE ! after reboot >>What are the differences between the working and the not-working cluster configurations? Both have exactly the same blade configuration except the storage systems. >>Are BGP or other routing protocols in use? No. We reply to the following prompts: * Do you want to configure dynamic ROUTED or GATED routing [NO]: The current configuration for the default route is: PERMANENT Type Destination Gateway PN 0.0.0.0 10.68.1.250 * Do you want to reconfigure a default route [YES]: Enter your Default Gateway host name or address: >>Any reason that you're not running OpenVMS V8.4 with TCP/IP V5.7; something a little more recent than V8.3-1H1? Application has not been tested on 8.4. The workaround is to run the TCPIP$CONFIG before rebooting each server. Noel
Dennis Handly
Acclaimed Contributor

Re: tcpip route could not be fixed

(This and your first post are pretty unreadable.  It would be nice if you could use Post Options > Edit Reply to add some paragraphs and newlines.)