"Monitor" error in cluster environemtn

[IFX_1]

Is there anyone know what is the reason behind the error below? How to fix it?

S7_$ mon cluster
%MONITOR-I-ESTABCON, establishing connection to remote node(s)...
%VPM-W-NOCONNECT, Unable to connect to remote node SICAM8
-MONITOR-W-NODEINIERR, error during node initialization
%MONITOR-I-CONT, continuing....

[Karl Rohwedder]

Have you read HELP/MESSAGE NOCONNECT?

Is the node reachable via TCPIP?

regards Kalle

[Ian Miller.]

What versions of VMS are running in your cluster?

[IFX_1]

I can connect to the other node using TCPIP. And I can connect also using DECNET.

[IFX_1]

The OpenVMS I'm using is - OpenVMS V7.2-1H1

[John Abbott_2]

Are you running DECnet ? DECnet over IP ?

$ REPLY/ENABLE on the local and SICAM8 node, look for any security alarms (assume they're enabled)

Check the uaf has a VPM$SERVER account on both systems.

Changed anything recently in sylogin ?

J.

[IFX_1]

I'm running DECNET-PLUS.
VPM$SERVER account exists in both nodes.

You are correct that security is enable. What could be the problem? There is no change done on sylogin.

[John Abbott_2]

Check on both systems the ncl start-up script NET$APPLICATION_STARTUP.NCL exists. Check the VPM session control information exists in the file, something like;

!* Session Control Application VPM added by
CREATE NODE 0 SESSION CONTROL APPLICATION VPM
SET NODE 0 SESSION CONTROL APPLICATION VPM ADDRESSES = {NUMBER=51} -
,CLIENT = -
,INCOMING ALIAS = TRUE -
,INCOMING PROXY = TRUE -
,OUTGOING ALIAS = FALSE -
,OUTGOING PROXY = TRUE -
,NODE SYNONYM = TRUE -
,IMAGE NAME = SYS$SYSTEM:VPM.EXE -
,USER NAME = "VPM$SERVER" -
,INCOMING OSI TSEL =

Are you using a cluster common system disk ? Are you using a common UAF file ?

[IFX_1]

The system disk is common to both node in the cluster. And the UAF is shared in each node.

[IFX_1]

John,
I check my file NET$APPLICATION_STARTUP.NCL and below info is already inside.

!* Session Control Application VPM added 15-JAN-2001 14:52:57.62 by SYSTEM
CREATE NODE 0 SESSION CONTROL APPLICATION VPM
SET NODE 0 SESSION CONTROL APPLICATION VPM ADDRESSES = {NUMBER=51} -
,CLIENT = -
,INCOMING ALIAS = TRUE -
,INCOMING PROXY = TRUE -
,OUTGOING ALIAS = FALSE -
,OUTGOING PROXY = TRUE -
,NODE SYNONYM = TRUE -
,IMAGE NAME = SYS$SYSTEM:VPM.EXE -
,USER NAME = "VPM$SERVER" -
,INCOMING OSI TSEL =

[John Abbott_2]

1. What happens when you issue a $ mon clu on node SICAM8 ?

2. $ reply/enable=security and look for security audit alarm/events when you issue the failing $ mon clu command, this may give you a hint to your problem. (assume you have alarm monitoring enabled for Logfailure: network ($ show audit will tell you))

3. You could also try $ana/aud/sin=time_just_before_mon_clu_cmd/full and look for any VPM$SERVER event failure - status field should give hint as to what's wrong.

4. post output from $ MC NCL SHOW SESSION CONTROL APPLICATION VPM ALL

Kind Regards
John.

[John Abbott_2]

Sorry, please also post

$ MC NCL SHOW IMPLEMENTATION

[IFX_1]

$ MC NCL SHOW SESSION CONTROL APPLICATION VPM ALL

Node 0 Session Control Application VPM
at 2006-04-28-18:37:39.252+08:00Iinf

Identifiers

Name = VPM

Status

Process Identifiers =
{
}
UID = A81D0C13-B1CA-11DA-83A7-AA00040020B7

Characteristics

Client =
Addresses =
{
number = 51
}
Outgoing Proxy = True
Incoming Proxy = True
Outgoing Alias = False
Incoming Alias = True
Node Synonym = True
Image Name = SYS$SYSTEM:VPM.EXE
User Name = "VPM$SERVER"
Incoming OSI TSEL =
OutgoingAlias Name =
Network Priority = 0

Counters

Creation Time = 2006-03-12-13:18:12.284+08:00Iinf

<><><><><><

$ MC NCL SHOW implementation

Node 0
at 2006-04-28-18:39:02.379+08:00Iinf

Characteristics

Implementation =
{
[
Name = OpenVMS AXP ,
Version = "V7.2-1H1"
] ,
[
Name = Compaq DECnet-Plus for OpenVMS ,
Version = "V7.2-1 ECO02 16-JUN-2000 14:11:13.30"
]
}

[Volker Halle]

Ronald,

welcome to the OpenVMS ITRC forum.

What happens, if you try $ MONI SYSTEM/NODE=SICAM8 ?

Will a new NET$SERVER.LOG file be created in the default directory of the VPM$SERVER account (typically SYS$SPECIFIC:[VPM$SERVER] on the 'remote' node, i.e. check SYS$SYSDEVICE:[SYS*.VPM$SERVER]) ?

Does $ DIR SICAM8:: work from the account, from where MONI CLUSTER is failing ?

Volker.

[John Abbott_2]

Good point Volker... Ronald, check your NET$SERVER.LOG files are not at version 32767

[Karl Rohwedder]

Regarding NET$SERVER.LOG:
When I issue a MONI CLUS, the VPM sever processes use TCPIP and will not create any NET$SERVER logfiles.

regards Kalle

[John Abbott_2]

I have 4 net$server.log files, but then I'm using pure DECnet (not over IP). Not sure what Ronald is doing ?

Ronald, I can't see anything wrong with your ncl scripts. Can you post anything else missing from my questions and Volker's reply ?

How does the vpm$server uaf record look ? don't suppose someones disuser'd it or changed the login access/time ?

Thanks
John.

[Volker Halle]

Ronald,

MONITOR CLUSTER or MONITOR/NODE=xxx can connect to the remote systems using either DECnet or TCPIP. For TCPIP, it requires that the remote system is running the VPM_SERVER process (started with SYS$STARTUP:VPM$STARTUP.COM and listening on port 5354 - check with TCPIP> SHOW DEV/PORT=5354).

MONITOR cluster REQUIRES, that the SCS nodename of the remote nodes and the DECnet or TCPIP node name (configured on the local node) exactly match. It obtains the names of the remote nodes from the cluster information and tries to connect to them using those names.

Volker.

[IFX_1]

Hi Guys,
Thank you for your help.
I resolved the issue by adding a proxy account on the UAF username that execute the MON CLUSTER.

It is okey now.

[Jan van den Ende]

Ronald,

the way to thamk those that helped you is outlined in:

http://forums1.itrc.hp.com/service/forums/helptips.do?#faq

Proost.

Have one on me (maybe at the Bootcamp in Nashua?)

jpe

[IFX_1]

Thanks.