Decnet issues

[vmsserbo]

Hi,
We have been receiving emails quite frequently from one of our nodes, idicating that decnet is down and Ip is up. Decnet seemes to be coming down around the same time. It recovers itself. HEfre is the email we get from the system. Any suggestions? The network people can't find any answers. We are running phase iv on a microvax 3100

***********************************************
From: "CCTT07::SYSTEM"@ctt.com [mailto:"CCTT07::SYSTEM"@ctt.com]
Sent: Thursday, May 04, 2006 11:36 PM
To: FNF - TEAM Systems Support
Subject: PITTSB has Network Problems

NET_NODES running from CCTT07 reports the following:

PITTSB was discovered having Network Problems at
2006-05-04 22:47:59.75, and this is still happening.
Currently DECnet is DOWN but IP is UP.
The IP Address for PITTSB is 170.88.9.212, and the router is 170.88.9.193.
Please investigate and resolve any problems!

CCTT07 Time: 4-MAY-2006 23:35:50.98
***********************************************

Thanks!

[Art Wiens]

You can look at various Decnet counters to get some clues:

$ MCR NCP
NCP> SHOW LINE xxx-0 COUNT
and SHOW CIRC xxx-0 COUNT, SHOW KNOWN NODE COUNT etc.

You can CLEAR the various counters to set a baseline and monitor from that point. Try and look at the Decnet counters just after you get your email.

HTH,
Art

[Volker Halle]

Miles,

please consider to close the duplicate entry:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1025039

You would need to find the automatic procedure (maybe called NET_NODES ?), which sends these mails and find out, what kind of checks this procedures executes to determine, that 'DECnet is DOWN'.

Are PITTSB and CCTT07 node names ? Which one is the MicroVAX 3100 ?

The check for DECnet on a Phase IV system, do:

$ MC NCP SHO EXE (should be on)
$ MC NCP SHOW KNO CIRC (should be on)

Volker.

[Arch_Muthiah]

Miles,

Apart from DECNet, what is the other IP network installed on your env. Is there TCPWare or Multinet?


Archunan

[Arch_Muthiah]

Miles,

There were some ethernet drivers came with old VAX/VMS system. I don't know what driver you have in your env, but due to the known probelms on those old ethernet drivers, DECnet used to trigger DECNet down and UP messages. I am not sure this may be the reason as I don't know your h/w and OS env.

Anyway it is good idea to look into the file which Volker mentioned to find at which instances the messages fired.


Archunan

[Robert Gezelter]

Miles,

First, these messages are not coming from a base OpenVMS component. So, the first order of business is to identify the source of the messages. The second task is to look at what test (code) is being used to determine that DECnet is down.

Depending on the check used, it is possible that system load (on either the system that the check is on, or another system in your network), or traffic at some intermediate point, is causing the problem.

More data is required.

- Bob Gezelter, http://www.rlgsc.com

[comarow]

It could well be that TCPIP and DECNET are using different lan adaptors. Are they?

In addition,there might be a router that is bouncing or a duplicate node on the net.
If it's a phase IV node, $show net will show your router.

For a more general look at your network,
Anal/system
show lan - will identify all the lan
adaptors in your network
and
show lan/count will show you all the
errors.

If your counters are clean, the problem could well be outside your box. You could start by swapping ports and cables.

[Steven Schweda]

Add my vote for first finding the source of
the e-mail messages, then figuring out why
they're being sent. So far, a network test
program gone berserk would explain all the
symptoms. (And, so far as I can tell, the
only symptom is a rash of e-mail messages
claiming that there's a network problem of
some kind somewhere.)