Michael,
I think that most of your performance problems will go away when you convert to linux.
We started a new poller system (exp_poller) that is parallel, but it still needs testing.
Anyway the current poller is host-parallel and interface-linear.
Javier
Meyers, Michael wrote:
What I did to work around this is to create multiple pollers that are five minutes apart. We are monitoring over 400 interfaces on one box, and the poller takes about 15 minutes to complete, so I have added (now) 4 pollers. i.e. poller1, poller2, poller3, poller4, etc... Poller1 starts at 12:00am, Poller2 at 12:05am, and so on. In your situation, you should only need two pollers.
I have compared the graphs to other graphs monitoring the same equipment and it is accurate. It would be nice if the poller process wasn't so linear. But it's a small problem that can be worked around right now with the application. Once that gets resolved and polling "in parallel" if you will, becomes available, we will be converting to jffnms exclusively (Javier, new feature request? :) ) We have tested it for everything and I am working to port our current JFFNMS version to Linux so we can also use RTT with smokeping (we are using the windows version now, everything on one box).
-----Original Message----- From: Anang Syarifudin [mailto:[EMAIL PROTECTED] Sent: Wed 7/28/2004 12:19 AM To: [EMAIL PROTECTED] Cc: Subject: [jffnms-users] Poller speed
Hi Javier, I'm a happy JFFNMS user i've been using it since it's early version 2 years ago. Recently I aware about it's performance since the box that running it hung frequently, I've tried to separate mysql on more powerful box, reduce interface to be monitored, and so on. Currently JFFNMS box never hung again, but I'm afraid that polling process aren't accurate because it always overlap between poller. I made a test, I stop crond daemon, than I check wheter polling process still running, I found that to completely finish polling process took more than 7 minutes, while polling process period is 5 minutes [EMAIL PROTECTED] etc]# ps auxw | grep php nobody 24056 0.0 3.2 10928 8228 ? S 11:10 0:02 /usr/bin/php -q poller.php 2 0 1 0 0 [EMAIL PROTECTED] etc]# date Wed Jul 28 12:17:26 WIT 2004 [EMAIL PROTECTED] etc]# ps auxw | grep php nobody 24056 0.0 3.2 10940 8240 ? S 11:10 0:02 /usr/bin/php -q poller.php 2 0 1 0 0 I only monitored 3 host with 72 interfaces, connection speed is not an issue. Do you have any suggestion to fix this, any comment are appreciate, and thanks for your great JFFNMS :) Anang ------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click _______________________________________________ jffnms-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jffnms-users
s===
-- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Javier Szyszlican, Project Leader, JFFNMS [EMAIL PROTECTED]
I hope JFFNMS or I were helpful to you, if you can, please donate at http://jffnms.org/donate
------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click _______________________________________________ jffnms-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jffnms-users
