[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

1332.0. "Left Hitchcock" by MARVIN::CCHAPMAN () Mon Aug 22 1994 11:47

I'm working on the WANrouter90 V1.2 project checking out the 'left hitchcock'
code, and have a couple of questions....

RESETS
======
The 3 different routers - 

	DECwanrouter90  - used to be called SPOTS
	DECwanrouter90 EW - used to be called 90S2
	DECwanrouter90 EI - the ISDN version

all claim to support the 'left hitchcock' RESET command in the software. 

How do I get HUBWATCH (V3.0.0) to send the reset command?

Is there any point in supporting the RESET command? SPOTS doesn't even have a
way of indicating that it supports RESET according to the 'left hitchcock'
specifiaction I have (revision 1.1.2.7 94/01/21), because it uses DEChub-90
minimal identification rather than the Generalized Reply used by the EW and EI.


IDENTIFICATION/sysoids
======================
The Identification stuff doesn't work quite right either. It's OK for
the DECwanrouter90 (SPOTS) - but for the EW and EI, I see the following...

HUBWATCH 'sees' the DECwanrouter EW when it is plugged into the Hub, and
displays the appropriate picture OK, however it complains about the
sysObjectID each time HUBWATCH polls...

(D) fp: slot2: received unknown module sysoid: 1.3.6.1.4.1.36.2.15.4.10.1

This is when the EW firmware is running. Once the software starts up, it sends
the same sysoid with 1.2 added onto the end because it's version 1.2
software (1.3.6.1.4.1.36.2.15.4.10.1.1.2 gets sent). HUBWATCH doesn't like
this either. 

The ISDN box does the same, except its sysiod is 1.3.6.1.4.1.36.2.15.4.9.1

Where are these sysoids defined? Are we sending the wrong numbers, or is
HUBWATCH wrong? Should the software version number be added onto the end of the
sysoid like this? I couldn't find anything in the spec about this.

Chris.
T.RTitleUserPersonal
Name
DateLines
1332.1Those 3 little words... "No can do."SLINK::HOODI'd rather be at the KennebecMon Aug 22 1994 13:4610
Hmmmm.  You want HUBwatch V3.0 to recognize devices that didn't exist when
we designed HUBwatch V3.0.  That is a problem.  :-)

As far as new variations of the device (with new sysObjectId's), they have
to be registered in the HUBwatch sources.  It's too late for 3.1, but it can
be added to V4.0.  Talk to me off-line about how we do this.

You cannot make HUBwatch V3.0 send RESET to the wanrouter 90.

Tom Hood
1332.2sysOid registration.NACAD::GALLAGHERMon Aug 22 1994 14:1932
The following OIDs are registered:

>dec                   OBJECT IDENTIFIER ::= { enterprises 36 }
>ema                   OBJECT IDENTIFIER ::= { dec 2 }
>sysObjectIds          OBJECT IDENTIFIER ::= { ema 15 }
>
>decWanRouter90EI      OBJECT IDENTIFIER ::= { routers 9 }
>    wr90EIversion1    OBJECT IDENTIFIER ::= { decWanRouter90EI 1 }
>
>decWanRouter90EW      OBJECT IDENTIFIER ::= { routers 10 }
>
>                      -- 2 port WAN Router.
>    wr90EWversion1    OBJECT IDENTIFIER ::= { decWanRouter90EW 1 }
>
>                      -- a DECwanrouter90EW configured to run as a 
>                      -- Single Line Leaf Router.
>    wr90EWversion2    OBJECT IDENTIFIER ::= { decWanRouter90EW 2 }
>
>                      -- a DECwanrouter90EW configured to run as a 
>                      -- Single Line Branch Router.
>    wr90EWversion3    OBJECT IDENTIFIER ::= { decWanRouter90EW 3 }

Contact DSSR::REGISTER if you have questions about OID registration.
You'll also need to coordinate this with Tom (.1).

>Where are these sysoids defined? Are we sending the wrong numbers, or is
>HUBWATCH wrong? Should the software version number be added onto the end of the
>sysoid like this? I couldn't find anything in the spec about this.

Defined by the registrar.  Probably.  No.  No.  What spec?    ;-)

							-Shawn
1332.3Resets.NACAD::GALLAGHERMon Aug 22 1994 14:3421
>How do I get HUBWATCH (V3.0.0) to send the reset command?

I don't know.

>Is there any point in supporting the RESET command? SPOTS doesn't even have a
>way of indicating that it supports RESET according to the 'left hitchcock'
>specifiaction I have (revision 1.1.2.7 94/01/21), because it uses DEChub-90
>minimal identification rather than the Generalized Reply used by the EW and EI.

The DEChub900 Hub Manager (aka MAM) implements a MIB called the "Chassis MIB".
The DEChub900 Chassis MIB contains objects which allow devices to be reset.
When set to the proper value, like say "reset(1)", the Hub Manager will try
to reset the device.  If the device supports CSNMP, the CSNMP protocol is
used to reset the device.  If the device supports LH, an LH reset command
is sent to the device.  If the device doesn't support the LH reset command,
then nothing happens, oh well.  If the device does support reset then it
resets itself and everyone's happy.

This is one reason to support the LH reset command.  There may be others.

							-Shawn
1332.4Left Hitchcock - It's not Alfred's brother.CUJO::HILLDan Hill-NetConsultant-Denver-553-3624Wed Oct 05 1994 18:4415
    I must be living a sheltered life.  I am not at all familiar with 'Left
    Hitchcock'.
    
    It is supposedly the Management protocol used to manage these:
    	-DECrepeater 90C
    	-DECrepeater 90T
    	-DECrepeater 90T+
    	-DECrepeater 90FL
    	-DECrepeater 90FA
    
    What is LH?
    How does the DECagent 90 use it to manage the above repeaters?
    
    Thanks,
    Dan
1332.5You're lucky to be sheltered from this...NETCAD::BRAGDONWed Oct 05 1994 19:107
    Left Hitchcock is an ASCII only, single master, request/response,
    combination datalink and application protocol.
    
    Actually, it is related to Afred, although you are correct in stating
    that 'LH' is not his brother.  I believe 'LH' got its name from the
    Hitchcock-like profiles of the Request '{' and Response '}' symbols.
    One of them would be 'LH' ?
1332.6It's named after Al himself.NETCAD::GALLAGHERWed Oct 05 1994 19:1831
Yeah, I guess you need to get out more   ;-)

Briefly, the Left Hitchcock protocol was invented for the DEChub90.  It's 
a simple ASCII-based protocol.  Commands are proceeded with the ASCII "{" 
character.  (If you squint at the "{" character you can see Alfred Hitchcock's
likeness in profile. ;-) 

The DECagent90 uses the LH protocol to manage local devices.  Typical 
commands are "{ 4 <cr>" for "who's in slot 4".   Additional commands set
and get repeater characteristics.

To manage remote devices, the DECagent90 uses a remote DECbridge90 to
proxy LH commands to repeaters in the same hub and the DECbridge90.

The LH protocol is still used in the DEChub900 to manage "legacy" devices
like the old repeaters.  In this case the Hub Management Agent Module (MAM)
sends LH commands to the repeaters.  A DECagent90 is not needed to manage
local LH devices.  The MAM acts as a proxy agent in this case:  receiving
SNMP messages, and possibly translating them to LH commands to manage
the LH devices.

The LH protocol was judged to be too primitive for a mid/high-end hub
like the DEChub900.  The Compressed-SNMP (CSNMP) protocol was invented
to replace it.  Most new hub modules, whether in the 90 or 900 form-factor,
use CSNMP.  Some rudimentary LH commands are used to indicate that the
module is capable of speaking CSNMP.

That's it for the history lesson.  If you'd like more detail, let me know 
and I'll mail you the LH protocol specification.

					-Shawn