Conference 7.286::fddi

1442.0. "NISCA PROTOCOL VERS MISMATCH RCVD" by TIMABS::BACHMANN (Ronny Bachmann - CS Basel Switzerland) Mon Sep 12 1994 17:36

    Hi all ,
    
    Some Idea out there for this result?
    
    
    Config 
    -------
    
    2 x Vaxserver 4000-500 VMS 6.1 with DSSI Disk and Ethernet AUI and DEFQA
    M7535-AD Interface .
    
    10 x WS Model 4000-90 128 MB PAGE/SWAP Local Ethernet AUI and FDDI with 
    DEFTA on Turbo Channel.
    
    Made all Setup following Release Note VMS 6.1 for FDDI and Installed
    CSCPATCH_1160 for Shadowing. 
    
    Server are Connected on FDDI via Chipcom Hub MIC connector and on FLoor
    another Chipcom Hub all WS via STP and CDDI to the Hub . The FDDI is a 
    local ring not to the outside world. 
    The Router on this Net is a CISCO 7000 connected via Ethernet to both 
    server all are none routing. The WS are also connected to Ethernet be-
    cause of booting not possible over CDDI. 
    ALL Line MNA-0 resp FCA-0 comes up also circuit no problem at all, also 
    there was traffic over both medium when look in the counter. The config 
    was running 2 days.
    
    Problem:
    
    The connection was brocken over FDDI and in the Errorlog was the
    following message:
    
    
    NI-SCA SUB-SYSTEM VERS MISMATCH RCVD
    
    STATUS 		00000000
    			00000000
    DATALINK UNIT	    0001
    DATALINK NAME	41514603
    			00000000
    			00000000
    			00000000
                                	DATALINK NAME= FQA1:
    
    REMOTE NODE 	41434603
    			00000000
    			00000000
    			00000000
    					REMOTE NODE = FCA
    
    REMOTE ADDR		000000AA
    
    			    003A
    					ETHERNET ADDR=AA-00-00-00-3A-00
    
    LOCAL ADDR 		000400AA
    			    0483
    
    					ETHERNET ADDR=AA-00-04-00-83-04
    
    ERROR CNT		    0001	1. ERROR OCCURRENCES THIS ENTRY
    
    UCB$W_ERRCNT	    008A	
    
    				138 ERRORS THIS UNIT
    
    ****************************** ENTRY 13. *****************************
    
   THIS ENTRYS HAPPENS A HALF HOUER UNTIL THE WHOLE LAVC CLUSTER CRASHED,
    AND THE SATELLITES WHERE NO MORE BOOTING , THEY HANG IN QUORUM LOST
    BLOCKING ACTIVITY. THE SOLUTION WAS A RESTORE FROM THE DSA6:
    ($1$DIA0,$1$DIA6) SYSTEM DISK A WEEK AGO. 
    I DONT KNOW WHAT GOING ON WITH THIS FDDI ANY HELP WILL GREAT.
    
    PS: LINE AND CIRUIT FROM FDDI ARE OFF NOW
    
    
    			REGARDS RONNY MCS BCO 
    
    				SWITZERLAND
    
    

    HI ALL,
    
    SORRY THE ERRORLOG ENTRY IS NOT 
    
    NISCA SUB-SYSTEM VERS MISMATCH  RCVD IT IS NISCA PROTOCOL VERS MISMATCH
    

Looks like a packet got corrupted.  The NISCA protocol version should be
identical as passed in SCS messages between two systems running the same exact
version of OpenVMS, so if a mismatch of this fixed value was detected, I'd
guess the packet must have gotten clobbered somehow, or maybe there's a CPU or
memory hardware problem.

Part of the error log entry seems to support this hypothesis:

>    REMOTE ADDR		000000AA
>    
>    			    003A
>    					ETHERNET ADDR=AA-00-00-00-3A-00

This sure doesn't look like a valid DECnet-style Ethernet address in
the form of AA-00-04-00-xx-yy!

It's interesting that a system disk restore was needed to fix this.

    Hi,
    
    rep. to 1442.2 
    
    The HW could not be defekt because without FDDI connected now
    everything works fine over the Ethernet.
    
    Secondly the Problem to boot the WS into the CLuster was no more
    possible, because the System Disk was corrupted so after restore this 
    Problem disappeared.
    
    Has anybody a idea how I can troubleshoot this Problem with the FDDI?
    
    
      			Regards Ronny
    

    Hallo all,
    
    The solution to this problem was , I have to replace the DEFQA-AD in
    one Vax4000-500 and then the Cluster work fine. 
    
    		regards ronny