Can I this is the flow of updation of container :
1) release_cached_resources(): cache_handler.c
2) cache_timer:start: loading cache
3) helper:cache_handler:  loaded
Am I right ?

Does "helper" provide the cache functionalities ?

Can we  implement the enterprise mib without help of CONTAINER ?
If yes, how the  interface will change ?


Sanjay


Dave Shield wrote:
2009/4/9 sanjaykumar <sanjay.ku...@globaledgesoft.com>:
  
But what happen when the "subsystem" data is updated ?
How agent table "Container" Data structure will be updated ?
    

That's up to the agent.

Typically, the agent will use a cache timeout to decide whether
to keep using the local copy of the data.   Once this timeout
has expired, the container data will be discarded.
   The next time a request comes in, the container will be
re-populated with a fresh copy of the current data
(which will then slowly get out-of-date until the timeout is
reached again).


  
What about the "memory Consumption", We are maintaining
the two data structure for the same DATA
    

In terms of the whole system - yes.
In terms of the SNMP agent, no.

The agent has acess to exactly one set of data - typically held
in the container data structure.   The whole point of holding
this local cache is that the agent *doesn't* have direct access
to the original data - otherwise there'd be no point.


  
what exactly meant by the "subsystem" ?
    

Whatever is responsible for handling the stuff that this MIB
is meant to be reporting.

   Remember that SNMP is all about *management*. It doesn't
do anything useful itself - it just reports on (and/or controls)
the workers, who actually do the real work.

For example, suppose you're looking at a Web-Server MIB,
providing information about the behaviour of your Web server(s).
Number of current connections, where they're coming in from,
what pages have been served recently, server configuration, etc.
Stuff like that.

  In that scenario, the "subsystem" would be the collection of
web server processes.

   The SNMP agent isn't involved in providing web pages, and the
web server(s) can work perfectly happily without any form of SNMP
involvement.   They hold precisely the information that they need
to do the job they're there to do.

  But if the SNMP agent is to report on the status of the web servers,
it must be able to get hold of this information - either from the web
servers themselves, or from some other source (e.g. analysing
log files).
   So yes - this might well end up with some duplication - with
the same information held by the HTTP daemon (so that it can
serve web pages) and the SNMP agent (so that it can report on
what the web servers are doing).

That's a price you pay for not putting too much of an unnecessary
management load on the web servers, and let them get on with
some real work :-)

Dave

  

------------------------------------------------------------------------------
This SF.net email is sponsored by:
High Quality Requirements in a Collaborative Environment.
Download a free trial of Rational Requirements Composer Now!
http://p.sf.net/sfu/www-ibm-com
_______________________________________________
Net-snmp-users mailing list
Net-snmp-users@lists.sourceforge.net
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users

Reply via email to