Re: creating custom mib

Hello Áki Hermann,

Le 08/04/2014 01:45, Áki Hermann Barkarson a écrit :
> 
> [...]
> 
> To learn more about this I want to create my own MIB and I have an
> enterprise OID (1513) that I can use.
> 
> So I send traps with SNMPv2-SMI::enterprises.1513.... and passing
> along some variable and this is where I'm struggling with syntax and
> general lack of understanding. I want to format my mib in the correct
> way..

I wrote mibs a few years ago, and I think I can give you some advice,
but probably I will forget things, and maybe I will be wrong on some
points. I hope others would correct me if needed.

> maybe it should be split down into one mib defining objects and
> then I define the notifications in another mib..

I believe it is not necessary to write two separate mibs.

> and and.. so many
> questions :) I can see various examples out there but I'd love to
> have a conversation with someone who actually knows this stuff.
> 
> 
> This is what I have so far:
> 
> 
> Advania-Traps-MIB DEFINITIONS ::= BEGIN

Ok.

> 
> IMPORTS
>    enterprises                                  FROM RFC1155-SMI

This is the old deprecated SMIv1 version of the SNMP syntax. You
should import enterprises from SNMPv2-SMI instead, and don't forget
the ';' at the end of the IMPORTS clause.

But also, you need to import the OBJECT-TYPE, DisplayString and
NOTIFICATION-TYPE that you use, and the MODULE-IDENTITY macro that
you ought to use. ;-)

IMPORTS
   MODULE-IDENTITY, OBJECT-TYPE, NOTIFICATION-TYPE,
   DisplayString, enterprises                   FROM SNMPv2-SMI;

> 
> advania                                OBJECT IDENTIFIER ::= { enterprises 1513 }

I think that advania could be defined to be the one and only
MODULE-IDENTITY of the mib, especially if this mib will be the
only one (or the main) mib under enterprises 1513.

Though, I prefer not taking the responsibility of this choice.
You may read the standard, and wait for other answers, before
deciding what to do.

<http://tools.ietf.org/html/rfc2578#section-5>
<http://tools.ietf.org/html/rfc2578#section-5.6>

> objects                 OBJECT IDENTIFIER ::= { advania 1 }
> notifications       OBJECT IDENTIFIER ::= { advania 2 }

Unless you are absolutely sure that the traps will never been generated
using SNMPv1 protocol but only SNMPv2c or SNMPv3 -- and even if you are
sure of that -- it would be highly desirable that you define a prefix
with a sub-id having the value zero for the notifications :

<http://tools.ietf.org/html/rfc2578#section-8.5>

Hence :

objects                 OBJECT IDENTIFIER ::= { advania 1 }
notifications           OBJECT IDENTIFIER ::= { advania 2 }
notificationsPrefix     OBJECT IDENTIFIER ::= { notifications 0 }


> 
> -- Objects
> 
> trapState OBJECT-TYPE
>     SYNTAX      DisplayString (SIZE (0..31))
>     MAX-ACCESS  read-write
>     STATUS      current
>     DESCRIPTION
>             "A brief description."
>     ::= { msg 1 }

  { objects 1 }  I suppose

> 
> shortMsg OBJECT-TYPE
>     SYNTAX      DisplayString (SIZE (0..31))
>     MAX-ACCESS  read-write
>     STATUS      current
>     DESCRIPTION
>             "A brief description."
>     ::= { msg 1 }

  { objects 2 }  I suppose

> 
> longMsg OBJECT-TYPE
>     SYNTAX      DisplayString (SIZE (0..1279))

I'm not sure whether it is allowed to redefine a DisplayString with
a maximum size greater than 255. What is the opinion of other readers
of this mailing list ?

>     MAX-ACCESS  read-write
>     STATUS      current
>     DESCRIPTION
>             "Detailed info."
>     ::= { msg 2 }

  { objects 3 }  I suppose

Also, if these objects are not actual objects than can be read
and written independently of traps, the MAX-ACCESS should
probably be accessible-for-notify and not read-write.

> 
> 
> -- Notifications
> 
> customTrap NOTIFICATION-TYPE
>        STATUS current
>        VARIABLES { trapState, shortMsg, longMsg }
>        DESCRIPTION "Custom trap alert."
>        ::= { notifications 1 }

  { notificationsPrefix 1 }

(see above, about the required zero subid)

> 
> END
> 
> Not sure if this is how it's meant to be formatted, any help would be appreciated..

I hope other answers will follow mine.

Best regards,
-- 
Olivier Miakinen


------------------------------------------------------------------------------
Put Bad Developers to Shame
Dominate Development with Jenkins Continuous Integration
Continuously Automate Build, Test & Deployment 
Start a new project now. Try Jenkins in the cloud.
http://p.sf.net/sfu/13600_Cloudbees
_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users