On Wed, 23 Aug 2006 12:10:41 +0100 Dave wrote:
DS> > Could this be an option in SNMPD.CONF ?

I have thought of that. It probably wouldn't be too hard to do.

DS> Ideally (IMO), this module should be written to use the cache_handler,
DS> rather than manipulating the cache itself.   That would allow the
DS> cache timeout, etc, to be controlled dynamically via the
DS> NET-SNMP-AGENT-MIB::nsCacheTable

It does use the cache_handler. Does it need to do something special to work
with nsCacheTable? Of course, someone setting the ifTable value > ~30 seconds
on a 32bit system w/a full gigabit link will get inaccurate values in the
64bit counters in the ifXTable. :-/

DS>    However, as I said, the inner working of the MfD framework are
DS> something of a mystery to me, and I'm reluctant to start fiddling with
DS> a mechanism where I don't really understand what's happening.

The ifTable is a special case, as I mentioned in my previous message. We need
to update the cache every 30 seconds to ensure 32bit counters don't wrap
before we poll again. This is documented in the comments where the cache
timeout value is set.

In theory, we could eschew the cache handler in favor of a simple 30 second
timer, reset every time the data was polled. This would handle the 32bit wrap,
while also allowing fresh data for every request.

But given that it's rare to poll every second (I think 5 minutes is probably
the most common), I'm not all that interested in doing the extra work.

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders

Reply via email to