I'm also working on a poller to collect a few thousand variables into an SQL
database. I haven't looked much at network efficency yet, mainly because it
seems to be dwarfed by the speed of updating the SQL DB. If memory serves
correctly it took less than 15 seconds to poll every interface (it takes 2
minutes for the entire script to run - the rest being mostly SQL).
I'd certainly be interested in seeing what you come up with though if you can
make it more efficient.
I have a vague memory of the script not exiting but the only thing I can think
of is the 'snmp wait' command???
I've put my project on hold for a few weeks. When I finish it (including docs
etc.), I do intend to post it as an example.
For the aspect of writing efficent programs, I tend to adhere to the principle
of doing this _after_ writing it and only if it seems to need it. Often as
not the things we think will need optimised are not where the bottlenecks
occur (IMHO of course). One thing that did occur to me to wonder about (in my
naive way) is whether a lot of the work can in fact be done whilst waiting for
packets to arrive (which is a _long_ time in terms of CPU cycles), anybody
have any comments on that?
I think some of the TCL extensins (such as tclX) do signal handling.
Eddie
-- !! This message is brought to you via the `tkined & scotty' mailing list. !! Please do not reply to this message to unsubscribe. To subscribe or !! unsubscribe, send a mail message to <tkined-request@ibr.cs.tu-bs.de>. !! See http://wwwsnmp.cs.utwente.nl/~schoenw/scotty/ for more information.
This archive was generated by hypermail 2b29 : Mon Jan 08 2001 - 15:27:35 MET