Video Screencast Help

NS Server Perfomance Issue

Created: 20 Sep 2013 | 11 comments
KNP's picture

Hi,

I have NS Server created and about 4000 agents are installed they are communicating proper, but after increasing number of agents i.e 5000 my NS Server creates a perfomance issue, even i have created package server for each site.but altiris console is opening very slow, i am not able to push agents through altiris as well i have observed that the NS Events are processing very slow and increases.

please suggest.

Comments 11 CommentsJump to latest comment

ianatkin's picture

First point of call is to client configuration request interval and collection update intervals. It might well be that these need to be scaled back to suit your increased environment.

For your collection update settings check here,
Configuration -> Server Settings -> Notification Server Settings -> Automatic Collection Updating

For your Altiris Agent settings (for configuration update frequency) check your policies here,
Configuration -> Altiris Agent -> Altiris Agent Configuration 

Your clients shouldn't really have to call back to the NS more frequently than every 4 hours I expect.

Post a screenshot (or link to one) if you're unsure whether what you are seeing is good/bad and we'll explain.

Also check taskmgr (cpu and ram) on your NS and SQL Server to check you haven't got a runaway process. Also check SQL server for general load issues.

Ian Atkin, IT Services, Oxford University, UK

Connect Etiquette: "Mark as Solution" those posts which resolve your problem, and give a thumbs up to useful comments, articles and downloads

KNP's picture

Hi ianatkin.

Our agent communication interval set as  6 Hours, so i think it is right configuration, as well i have checked task manager CPU and ram usage it is using both uses 50% usage as well Network usage is 14% Link is 1 GBPS.

please suggest further steps to increase perfomance..

Sachin Sawant's picture

Hi check the NSE file status if it’s not posses properly please follow-up following link.

 

http://www.symantec.com/connect/forums/problem-nse-files

ianatkin's picture

Hi,

6hrs for client configuration updates look good. You haven't mentioned what your collection update intervals are though??

 

Ian Atkin, IT Services, Oxford University, UK

Connect Etiquette: "Mark as Solution" those posts which resolve your problem, and give a thumbs up to useful comments, articles and downloads

RRonlev's picture

KNP,

While I appreciate everyone's comments on "stuff to check", in my opinion performance issues can arise from all manner of configuration settings (issues) as well as (hardware) bottlenecks both on the SMP as well as SQL server predominantly.

Adding the additional endpoints may very well have tipped the scale - or just exposed a configuration issue, say based on new sites not handled by existing site assignments.

In general when the Console slows down - I would look at a couple of things;

1.) SQL/CMDB; What's the fragmentation of the various tables? And do you have a maintenance plan in place.

2.) Parse the IIS logs (if you know how to) and see what page hits are taking place. You may not really have a good baseline, but if you see a lot of task server registrations that would be bad (within short time intervals)

3.) Review the NSE queues (as suggested) as well as reoccuring policies and/or tasks, that might create data. While this may (or may not) show up as CPU utilization it very much will show up as disk I/O on both the SMP and SQL server, so running perfmon on the two might also give you an idea of whether you happened to hit some issue like that.

Last but not least, what is/was your hardware spec'ed out to support, when you add more endpoints your hardware demands are increased, so compare against the installation guide for that too.

For some reference material check out http://www.symantec.com/docs/HOWTO10330 in the Performance Tuning section.

Regards,
Rasmus

ianatkin's picture

RRonlev,

You are of course right, when looking at performance there are many metrics to look at and examine.

In this situation where someone increases their client base by 25% and goes from good performance to really bad performance instantly it suggests (to me) that perhaps something else has changed in the environment. And some admins can fiddle with settings the wrong way when increasing their environment, most notably collection updates frequencies and client config updates.

So, I always check these obvious things first (which take seconds to check) before delving into the much more more time consuming IIS logs, SQL logs, package server configurations,  etc.

It's just my approach troubleshooting approach.

And, my next port of call here (if KNP is still watching this thread??!) would be to ask for package server configuration details and where the backend SQL is. Again, (hopefully) fast and easy to get info. My gut feeling (which doesn't mean I'm right!) is that the newly introduced 1000 clients are hammering the server with package downloads.

Your reference docs are however entirely valid and should certainly looked at.

Kind Regards,
Ian./

Ian Atkin, IT Services, Oxford University, UK

Connect Etiquette: "Mark as Solution" those posts which resolve your problem, and give a thumbs up to useful comments, articles and downloads

KNP's picture

Hello,

 

@ianatkin my other collections on weekly basis shceduled.

ianatkin's picture

Hi,

Well, DEFINITELY not that then (though collections updates every two weeks is a config I've not come across!).

Further, after two weeks (that's how long it's been since you opened this thread) I would have assumed that initial excess of package downloads (if package servers had been misconfigured) would have sorted itself out.

I would suggest then you follow the other good advice in this thread to isolate what exactly is causing the problem. Keep the thread updated if you want with progress, what you've checked etc... and then people can chip in.

A tip for something to check outside of the NS application. If you working off virtualised kit, make sure someone hasn't done a storage migration on you or reduced your VM resources.

Kind Regards,
Ian./

 

Ian Atkin, IT Services, Oxford University, UK

Connect Etiquette: "Mark as Solution" those posts which resolve your problem, and give a thumbs up to useful comments, articles and downloads

RRonlev's picture

KNP,

 

Just an off the cuff thought; Are you having issues with pushing the agent on the "agent push page" in particular? I.e. long page loads or timeouts on that specific page? I've come across that issue before.

 

Regards,

Rasmus

KNP's picture

Hello RRonlev,

 

I am also facing issues with Pushing the agent from altiris Console.

 

 

HighTower's picture

KNP, have you resolved your performance issues?

If not, can you provide more information about the version of the platform you're running, your architecture (server, SQL server, etc.), and an update of what you've tried so far?