On Fri, 24 Feb 2006, David T. Perkins wrote:
> HI,
>
> I'm concerned a little over the use of different terms
> that mean essentially the same thing, since this may confuse
> many people. That is, I've seen "alarm", "alert", etc.
>
> In all cases, it appears to me, that you are trying
> to find an efficient coding technique to determine
> when to generate an "event report" (which is called
> a notification in SNMP).
OK, what I wrote was "what is basically an alarm". This is because I am
really an Internet guy who for the last seven years have been employed by
the telecom world (Ericsson) - that colors my language somewhat :-) In
the telecom world an "alarm" is (roughly) something that can be reset - in
that language the ifUp/ifDown pair constitutes an "alarm". And that is
what we are talking about here.
> In general, code can poll
> (repeatably query) a "resource manager" for status
> or statistics to determine if an event has occured,
> or "subscribe" to a resource manager to be notificated
> via a message or callback when something of interest
> occurs.
Here is where I find that the concept of an "alarm" is useful: The
polling technique only works *because* the ifUp/ifDown pair is an "alarm":
A binary status in the managed system (oops, that was another telecom
term...), and the purpose of the exercise is to mirror that binary status
to the NMS, as efficient, quickly and trustworthy as possible. The
polling techique may be efficient (it isn't in the case of the originator
of this thread); it is neither quick nor fully trustworthy (it can
overlook a status change that is more short-lived than the polling time).
But it is trustworthy to the extent, that a long-living true status in the
managed system will be mirrored as a long-living status in the NMS.
> If the resource manager does not support
> subscriptions, or provide other mechanisms for
> independent processes to determine the occurance
> of events, then the implementation is stuck with
> polling. In general, in such a situation, I suggest
> that the developers of the resource manager
> be requested to extend it instead of doing
> a quick and dirty patch (which typically
> results in the functionality provided, but
> at a high cost in resources to the system).
Yes, but in this situation the "resource manager" (the linux kernel) does
"support subscriptions". The mechanism is called netlink! Modifying
net-snmp to employ that support instead of polling, thus gaining lower
load on managed system, immediate event issuing and no loss of short-lived
link-downs or link-ups - that cannot possibly be described as "a quick and
dirty patch".
best regards
--
Peder Chr. Nørgaard Senior System Developer, M. Sc.
Ericsson Denmark A/S, Telebit Division
Skanderborgvej 232 tel: +45 30 91 84 31
DK-8260 Viby J, Denmark fax: +45 89 38 51 01
e-mail: [EMAIL PROTECTED]
(old e-mail 2000-2003: [EMAIL PROTECTED])
(old e-mail 1992-2000: [EMAIL PROTECTED])
----------------------------------------------------------------
This communication is confidential and intended solely for the
addressee(s). Any unauthorized review, use, disclosure or distribution
is prohibited. If you believe this message has been sent to you in
error, please notify the sender by replying to this transmission and
delete the message without disclosing it. Thank you.
E-mail including attachments is susceptible to data corruption,
interruption, unauthorized amendment, tampering and viruses, and we
only send and receive e-mails on the basis that we are not liable for
any such corruption, interception, amendment, tampering or viruses or
any consequences thereof.
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
