topic Re: network partner logical link aborted in Operating System - OpenVMS

network partner logical link aborted

Maurille Emmanuel — Tue, 11 Oct 2005 05:26:47 GMT

Hye,

I have updated my openvms 7.3.2 system to the eco2. And the following problem appears when I try to go to on other user session. The following message appears. How can I resolve this problem?

Trace :
------
set host 0
%SYSTEM-F-LINKABORT, network partner aborted logical link

Thanks for all your answers.

cdt

Re: network partner logical link aborted

faris_3 — Tue, 11 Oct 2005 06:55:26 GMT

Hi,
There could be many different causes.
Some articles on the topic:

[DECnet/OSI] SET HOST 0 Returns SYSTEM-F-LINKABORT

Copyright (c) Compaq Computer Corporation 2000. All rights reserved.

PRODUCT: Compaq DECnet-Plus for OpenVMS, Version 7.1

COMPONENT: NET$ACP

OP/SYS: Compaq OpenVMS Alpha, Version 7.1-2
Compaq OpenVMS VAX, Version 7.1-2

SOURCE: Compaq Computer Corporation

SYMPTOM:

SET HOST node_name works as expected, however, issuing a SET HOST 0 returns
the following error:

%SYSTEM-F-LINKABORT, network partner aborted logical link

Examining the NET$ACP process quota using "SDA> SHOW PROC NET$ACP/PHD"
shows the Free Page File Pages to be at or near zero, and the network
cannot be restarted.

COMPAQ RESPONSE:

Engineering has acknowledged this problem and plans to address it in a
future release.

WORKAROUND 1:

Contact Compaq Services for a workaround to this problem.

WORKAROUND 2:

Increase PGFLQUO for NET$ACP process via NET$LOGICALS, minimize DECnet
connection rejections.

ANALYSIS:

DECnet-Plus was brought up with system default quotas, which has 2048
PGFLQUO for NET$ACP. Over a period of two to three weeks, Free Pagefile
Pages for NET$ACP nears zero and DECnet becomes unusable.

This condition appears to be triggered by incoming DECnet connection
rejections, where an object has not been defined and external nodes are
attempting to connect to this object (i.e., SNS$WATCHDOG):

%%%%%%%%%%% OPCOM 26-MAY-2000 16:25:31.09 %%%%%%%%%%%
Message from user SYSTEM on ASPEN
Event: Access Control Violation from: Node LOCAL:.ASPEN Session
Control, at: 2000-05-26-16:25:31.092+00:00Iinf
NSAP Address=39:840:11-35-67-00-1B:AA-00-04-00-D3-6C:20,
Source=name = SYSTEM,
Destination=name = SNS$WATCHDOG,
Destination User="",
Destination Account="",
Node Name=
eventUid 40E00CE8-3322-11D4-802E-08002BBCF1E5
entityUid E2FBF0F7-2830-11D4-8001-AA000400935C
streamUid 0C2B0C4C-2831-11D4-8002-AA000400935C

Stopping the external connection attempts (in this case SNS) causes the
Free Pagefile Pages for NET$ACP to no longer decrement.

RELATED ARTICLES:

Other articles in the DECNET-VMS database discuss this error and SET HOST
0. These articles may be found using a search string of:

SET HOST 0 SYSTEM-F-LINKABORT

Re: network partner logical link aborted

faris_3 — Tue, 11 Oct 2005 08:29:01 GMT

Does

$ set host x.y

work ?

If x.y is your Decnet Address.

Re: network partner logical link aborted

Maurille Emmanuel — Tue, 11 Oct 2005 09:13:53 GMT

Homi

Yes the following command is working
set host x.y

Emmanuel

Re: network partner logical link aborted

faris_3 — Tue, 11 Oct 2005 09:45:24 GMT

Emmanuel,

try a
$ delete sys$system:dns$cache*.*;*

then
$ mc ncl disa dns clerk
$ mc ncl ena dns clerk

Re: network partner logical link aborted

faris_3 — Tue, 11 Oct 2005 09:55:44 GMT

Did you install
HP DECnet-Plus for_OpenVMS_Alpha Release Notes for Version 7.3-2

ECO01 or ECO02 ?

In the release notes of ECO01 you can read :

This kit contains a correction to the CDI$SHR
executable (see CDI$SHR.EXE in the Corrections
section). To ensure that this correction takes effect,
use the following command to flush the CDI cache after
you install this ECO kit:

$ MCR NCL FLUSH SESSION CONTROL NAMING CACHE ENTRY "*"

Have you tried flushing the cache ?

Re: network partner logical link aborted

Michael Yu_3 — Tue, 11 Oct 2005 18:49:45 GMT

Hi Emmanuel,

Set host 0 did not work but set host x.y (phase IV address of local node worked).

By default, set host 0 uses OSI transport while set host x.y uses NSP transport.

Set host 0 also depends on the in memory CDI cache for outgoing connection. If the in memory CDI cache is not valid, then it may have this problem.

I think the problem was introduced when the in memory CDI cache was updated (say, when node addressing was modified), but system was rebooted before the on disk CDI cache was updated.

The following NCL should fix the problem.

NCL> flush session control naming cache entry "*"

Thanks and regards.

Michael

Re: network partner logical link aborted

Maurille Emmanuel — Wed, 12 Oct 2005 03:20:42 GMT

Hye,

1- The flushing naming cahe entry command has been done.
2- I tryed the Homi solution but it dos not work.

Have you other idea.

The eco version is :

HP TCP/IP Services for OpenVMS Alpha Version V5.4 - ECO 2
on a AlphaServer ES40 running OpenVMS V7.3-2

Re: network partner logical link aborted

faris_3 — Wed, 12 Oct 2005 04:23:52 GMT

Yes. Open a call with the French CSC :-)

Re: network partner logical link aborted

faris_3 — Wed, 12 Oct 2005 09:01:14 GMT

For information,

the problem was solved by installing
HP DECnet-Plus for_OpenVMS_Alpha Version 7.3-2 ECO01

ECO01:

- Corrected a problem with Domain proxies brought
about by using Session Control V3 (SC3) with RFC1006
connections. The problem was corrected by reverting
IP address towers back to SC2 (see SYS$LOADABLE_
IMAGES:NET$SESSION_CONTROL.EXE corrections). CDI
is now used to look up the node name for RFC1006
connections. This can cause rejected connections if
the backtranslation entries in DECdns are incorrect. To
ensure that this change takes effect, use the following
command to flush the CDI cache after you install this
ECO kit:

$ MCR NCL FLUSH SESSION CONTROL NAMING CACHE ENTRY "*"

(ftp://ftp.itrc.hp.com/openvms_patches/alpha/V7.3-2/AXP_DNVOSIECO01-V732.txt)