Conference 7.286::fddi

2073.0. "SFT III Server Switching with DEFPA-UA" by TBJVOA::FUKUHARA_S () Mon Jun 24 1996 03:21

 Hello,

 I have a question.

 I tested for the switching of primary server on NetWare 4.1J SFT III
(the Japanease version) using DEFPA-UAs as MSL boards, and I don't
think one of the results was seemed proper.
 When the LAN cable of the primary server was pulled out, IPX root
seemed to be cut. It became impossible for the client to attach to
the server, or when the client had already logged in, the network
drive became unavailable.
 About the secondary server, it didn't have any problem --- whenever
the secondary server's  LAN  cable was drawn out,  the client could 
attach to it or network drive was still valid.

 I'd like to know if the phenomenon is adequate. Or if it isn't so,
what is the problem --- the test environment, the "set" parameters,
SFT III of Japanease version, or MSL driver --- ?

The test environment:
    Servers
        PCs:          DEC Prioris XL Server 560       x 2
        Memories:     24M    x 2
        MSL Boards:   DEFPA-UA (A02)    x 2
            Driver:   DECMSL4X.MSL (V1.20 1995/11/10)
        LAN Boards:   DE500-XA(B01)     x 2
            Driver:   DE500.LAN  (V2.92  1996/05/31)
                              or
                      DE435-AA(B01)     X 2
                      DE43X.LAN  (V2.60  1995/11/10)
        "set" command parameters:   default

    Client
        PC:           Celebris 5100
        Memory:       24M
        LAN Board:    DE500-XA(A01)
            Driver:   DE500.COM  (V2.10  1996/05/30)

I changed the LAN board on each server, and tested different boards
and drivers. So, the LAN boards don't seem that they cause it.
I searched for the topic about this question, but I couldn't find.
And I don't know which conference would be the best to ask this
question. 

Any reply will be helpful --- such as  "You should see the conference
XXXX::XXXXX", "You should see number XXXX.XXXX" or "Your test was
wrong because..." etc.

Thanks.
 

    
    
    
    
    
>> When the LAN cable of the primary server was pulled out, IPX root
>>seemed to be cut. It became impossible for the client to attach to
>>the server, or when the client had already logged in, the network
>>drive became unavailable.

This appears to be consistent correct behavior.

When the primary cable is pulled, the server has no idea where the
connection is broken, so the clients will lose connection.  If the
primary goes completely down, then the secondary will realize it
over both the LAN and the MSL connection.  The secondary will then
send a router configuration update message to the clients, and take 
over.

>>I searched for the topic about this question, but I couldn't find.
>>And I don't know which conference would be the best to ask this
>>question. 

Actually, if you have more questions on this topic, you might try
contacting technical support at Novell.  I'm not sure that this is
"correct" behavior, but it's consistent with what we have seen in 
the past.  You might try asking Novell if they are aware of this
situation.

-Brenda
 

-Sachiko