From: Jeff Allen <jra@corp.webtv.net>
Berg, Ivan wrote:
> Actually the problem happens on a couple of routers just when they get
> reset. The two that it has already happened to is two cisco routers whose
> power reset to clear problems.
This is a dead horse. Flog it all you want, but it comes down to this:
a. solving the negative spike problem requires fetching
the time at which the coutner was reset (the sysUpTime, or
the goofy Cisco-specific coutner reset time) along with the
data.
b. most people are happier with the rrd-max solution than
fetching extra stuff and trying to make sense of it on the
NMS side.
c. Cricket does not support solution (a), and I will not
be taking the time to implement it. If you want to
implement it, and you can do it in a way that makes it
optional for those of us who don't want it, then I'll
happily integrate your patches.
> Like I said, it wouldn't seem to be too hard to detect a negative
> change.
It's a harder problem to solve well than you think it is. If you don't
believe me, implement it yourself and send me the patches. Then I'll
be proven wrong, and Cricket will be all the better for it. :)
-- Jeff R. Allen | jra@corp.webtv.net WebTV Service Engineering | http://www.munitions.com/~jra--------------------------- ONElist Sponsor ----------------------------
Start a new ONElist list & you can WIN great prizes! http://www.onelist.com See homepage for details on ONElist’s new "FRIENDS & FAMILY" program.
------------------------------------------------------------------------
This archive was generated by hypermail 2b29 : Mon Mar 06 2000 - 19:00:56 PST