[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference netcad::hub_mgnt

Title:DEChub/HUBwatch/PROBEwatch CONFERENCE
Notice:Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7
Moderator:NETCAD::COLELLADT
Created:Wed Nov 13 1991
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:4455
Total number of notes:16761

4196.0. "900TP Port Driver Problems??!!" by ZUR01::ACKERMANN () Mon Feb 03 1997 11:59

    
            !!!"Portswitch 900 Quality Problem ? !!!
    
    
    Hi together,
    
    Big Customer is complaining about our Portswitch 900TP.
    He says,that intermittend PCs dont work on different Ports. He replaced
    Cabling and PCs, but some Ports just dont work. This happens on
    different Portswitches on differnet Ports. He replaced also the
    Modules twice, but Problems remains.
    Now he send me some of the Modules. I tested all the Ports and
    really, there are Ports, they dont get status online. The Ports are
    all default in Group one, and Thinwire Connected. (Status after Factory
    Reset)
    I change the Cable to another Port, this works. Back to to the
    other Port, dont work. Now i disable "Link Test Admin Status" on
    this Port, and now it gets online and works! (Without unplug Port!)
    This happens on different Modules, different FW Revisons, actual
    V2.1.1
    So, what does "Link Test Admin Status disable" change on the Port?
    This can not be the Solution, because when disabled, the Port LED
    ist always online, if selected, You have also no Status Information
    of the Module.
    Is this a known Problem, or Bug, or Problem of Drivers on the
    Port ?
    Any Help is wellcome,
    
    Thanks,
    
    Daniel MCS Switzerland
    
T.RTitleUserPersonal
Name
DateLines
4196.1Not an answer, just a data point.NETCAD::GALLAGHERMon Feb 03 1997 12:3541
Daniel,

Here's what the MIB says about "Link Test".

>erptrMauLinkTestAdminStatus OBJECT-TYPE
>    SYNTAX  INTEGER {
>               enabled(1),
>               disabled(2)
>            }
>    ACCESS  read-write                                -- ** Non-volatile
>    STATUS  mandatory
>    DESCRIPTION
>            "Link test functions detect certain types of media faults on
>            point-to-point link segments connecting pairs of link-type MAUs
>            (e.g. 10BASE-T, 10BASE-F, FOIRL). This object controls the link
>            test function implemented by some link-type MAUs.
>
>            Setting this object to enabled(1) causes the MAU to enable its
>            link test function.
>
>            Setting this object to disabled(2) causes the MAU to disable
>            its link test function. Disabling a MAU's link test function
>            can be useful for compatibility to link-type MAUs which do not
>            implement a link test function.
>
>            Link-type MAUs which implement link test functions but do not
>            support software control of the functions allow this object to
>            be set to enabled(1).  A badValue error is returned if this
>            object is set to any other value.
>
>            Coax type MAUs (e.g. 10BASE5, 10BASE2) or any other MAUs which
>            do not implement link test functions return disabled(2) when
>            this object is read. These MAUs allow this object to be set to
>            disabled(2) but return a badValue error if the object is set to
>            any other value.
>
>            Note: The state of this object is preserved when the repeater 
>                  undergoes a power-up reset."
>    REFERENCE
>            "Reference IEEE 802.3 Std., sections 9.9.2.1 and 14.2.1.7."
>    ::= { erptrMauEntry 4 }
4196.2Link Integrity issue for sureNETCAD::GILLISMon Feb 03 1997 14:4540
Daniel,

Shawn nailed it. What DEC calls Link Test Admin Status is a
software interface to the low-level hardware Link Integrity feature.

Link Integrity (when enabled on a port or a MAU) is a test pulse
sent out the Tx pair of a repeater port. (I'm talking at the repeater
chip level; what the user sees on the front panel are just media
interface ports, such as ones that accomodate 8-pin modular jacks).
If the pulse is not received on the Rx pair after a short delay,
indicating a link integrity failure, the repeater chip
STOPS REPEATING DATA TO THAT PORT. The port LED on the front panel
also gets shut off (the port LED is an indication of link integrity,
not of port state).

Link integrity MUST be set to the same state (enabled or disabled)
on both sides of a link. Old computing devices and some really old
MAUs and repeaters don't support link integrity, so link integrity
is seen as disabled. 

It sounds like your PCs may have fallen into the latter category.
Perhaps your MAUs that attached the 10BaseT cable to your PC's
NIC cards have link integrity disabled, or do not have the setting
available. (Some MAUs have internal jumpers, others have DIP switches
or "slider" switches).

Also, for those ports which you said worked, what was the state
of Link Test Admin Status for those repeater ports?

Link Integrity is one of the primary things network managers define
RMON alarms on. If link integrity changes state on any repeater
port in a topology, it's lets you know that a device has failed,
someone has unplugged something, or sabotage has occurred.
Link Integrity is a network manager's friend, and should only
be disabled when the device on the other side of the link
doesn't support the feature (for backwards compatibility to
older devices which don't support link integrity).

John Gillis
clearVISN
4196.3Thanks and some more Infos..ZUR01::ACKERMANNTue Feb 04 1997 09:5022
    
    Hello together,
    
    Thanks for Your detailed Explanations.
    
    >Also, for those ports which you said worked, what was the state
    >of Link Test Admin Status for those repeater ports?
    
    The state was enabled. Strange is, that most of the Ports work
    correctly. If i change Link Test Admin Status to disable, than
    enable again, unlug, then plug in again the same Cable with
    nothing changed then it works. It just needs a "kick" to get
    running. Maybe there is a low threshhold on the Signal, that 
    causes this go -nogo effekt.
    But we go now Onsite to Customer Site to check, wheter the
    Symptoms are exactly like we had it here in our Lab.
    Maybe somebody else have had similar Symptoms bevore?
    
    Thanks,
    Daniel MCS Switzerland