Dynamics 365 F&O/AX Dynamics 365 CE / CRM Dynamics GP Business Central / NAV UG Dynamics SL Power Platform Fabric UG MS Copilot UG

[Karli Mather]

Hello,

Is there a correlation between enabled license codes and performance? My organization is fairly new to AX and has many codes enabled for modules we don’t use. We have fewer than 100 users with two AOSs for users and one for batch processes and we are already having performance issues (tasks like clicking on menu items taking much longer than they should). Additionally it appears that our load balancing is following a stepwise pattern rather than true distribution of users and processes. Any suggestions would be greatly appreciated. Thanks!

——————————
Karli Mather
Fidelitone
Wauconda IL
——————————

[Darrell Smith]

Karli,

I agree with Jackie that there are many reasons for performance degradation and she gives a lot of good starting places.  To answer your specific question, having 2 AOSs, and a separate batch server should be sufficient to support 100 users.  That assumes they are sized and configured properly.  I didn’t see what release you were at but I assume AX2012.  Also, having license keys enabled for modules/areas you don’t use has little affect on performance and virtually no affect whatsoever on menu items being slow.

Good luck with your tuning.  It’s always a journey.

Darrell

——————————
Darrell Smith
Application Consultant
Agility Business Solutions
Cleveland OH
————————————————————————-

[Luis Xavier Baena Mourão]

Hi Karli,

Darrel has given you the answer. The affect on performance of having functionality enabled that you don’t use is negligeable. However it’s a best practice to turn off whatever you don’t use.

Best regards.

——————————
Luis Xavier Baena Mourão
Technical Expert
Axllence Schweiz
Wallisellen
————————————————————————-

[Jack Moran]

Hi Karli,

AXUG folks as always offering lots of good advice on your question. Think you can readily see that researching and addressing performance issues in an AX environment is a complicated matter. You could spend days/weeks trying out different theories only to find yourself in the same predicament.

So my suggestion for you is to get acquainted with someone who specializes in this area and seek their guidance. Specifically suggest you check out this link: http://csgax.com/services/revive/ax-performance-tuning. 

——————————
Jack Moran
Director of Industry Solutions
Columbus
Greenfield NH
————————————————————————-

[Thomas Bonde Ejby]

Hi all Just wanted to comment on this, as performance always is a topic for discussions between technical guys, developers and consultants
And as said this is an never ending journey with a lot of variables. But to answer your initial question, the answers is YES, configuration keys can have an enormes impact on performance. And here is the reason why.

• Configuration keys can be associated with Index and code, hence turning of and on bad code or bad index
• Configuration keys can therefore open up code and fields that HAS to be filled out in order for the application to work proper. I have many examples on this, one was a installation in PERU where the Swedish Country key was activated, and because the obviously did not configure the fields for Sweden the payment journals was hit with a performance degrade factor in the hundreds
• Another in AX2012 R3 we have seen that if you have Call Center turned on under Retail, but not proper configure, the standard Invoice process takes a hit

The problem is that there is not cookbook that tells you what to do because it depends on how you use the system, how many data etc. For AX more is not better, be very careful about this, if you have 10 persons would it be best with 1 or 10 AOS. Because AX uses Caching mechanism your performance will be very bad with on user on one AOS instead of 10 users on 1 AOS. You have to design your system and monitor your performance according to you current situation

There are many tools, traceparse, debug, Dynamics Perf and others, but they are no good if you don’t understand the result and you don’t understand your business, this is why sometimes it is difficult for an outside to get the grip of what is going on.

I would be happy to dive deeper into this discussion with you, and am keeping AXUG session around the world on the topic, the next will be in Chapter meeting in Redmond, and Portland and then in at AXUG Summit EMEA.

So back to you direct question, does AX Config keys and license keys have an impact on performance. The answer is YES, and remember Microsoft supports you to add keys and license but actually they don’t really support you to take them away, as you could have code related to the keys that would not work.

Strip down your application to what you need to avoid any potential problems in your application, and then consider and test carefully when you turn keys back on.

And to your next question, how do a measure ? My answer will be baseline your system. It is not an easy task but you should take you application. And run through you vital process, and try to baseline how long time it takes to invoice, MRP plan etc. together with some technical baseline of your system I/O, network etc. This is the only way you can measure if any changes has happen after an update, code change or simply because you have more and more data in your application and the SQL server start behaving differently just because of the amount of data

Hope this give you an inside and your are welcome to give me a call, to talk further and if you would like I can try to arrange an online session through AX UG about this “Configuration key” topic

/Thomas TEJBY@axvice.dk

——————————
Thomas Bonde Ejby
Enterprise Solution Architect
Dynamics AX MVP
AXVICE IVS
——————————
——————————————-

[Jackie Olson]

Hi Karli,

Are you seeing this performance issue in all your AX environments?  (DEV, TEST, UAT, PROD)

Performance can be cause by lots of different things – hardware (disk), memory (lack of), bad code, bad/missing sql index, model db that is not optimize, network etc.

Did this issue just started recently? Have you tried using the AX trace tool to see what is causing the bad performance?

Do you have custom AX roles – is performance better for sys admin vs a user with custom AX role?

Were there lots of code change done recently?  If so, try running AXUtil Optimize.  More information can be found here – How to: Optimize a Model Store for Runtime [AX 2012]

Microsoft		remove preview
How to: Optimize a Model Store for Runtime [AX 2012] In versions of Microsoft Dynamics AX 2012 prior to Microsoft Dynamics AX 2012 R2, the model store tables are stored in the same database as the business data. If your environment contains a large combined business and model store database, we recommend that you avoid running the Microsoft Dynamics AX optimization command. View this on Microsoft >

I’ve seen performance issue with AIF when config keys were disabled – due to creation of Temp tables – Temporary TempDB Tables [AX 2012]

Microsoft		remove preview
Temporary TempDB Tables [AX 2012] In Microsoft Dynamics AX, one type of temporary table is a TempDB table. We call them TempDB tables because their TableType property value is TempDB. This value comes from the TableType::TempDB enum value. The TableType property value can be set at AOT > Data Dictionary > Tables > MyTempDBTable > Properties > TableType. View this on Microsoft >

Check out this blog for additional perf resources – Dynamics Ax Performance Team Blog

Microsoft		remove preview
Dynamics Ax Performance Team Blog We are happy to announce that the following benchmarks will be available, just in time for Tech conference 2015. We have published both summary and detail documents for each benchmark. Microsoft Dynamics AX 2012 R2 Benchmark for Multiple Retail Workloads In 2014, Microsoft conducted a benchmark on Microsoft Dynamics AX 2012 to measure the performance… View this on Microsoft >

For the load balancing – are you using a dedicated AOS load balancer?  If not, do you have both the AOS listed in the client’s AX client config?

Regards,

——————————
Jackie Olson
Stoneridge Software Inc.
Barnesville MN
————————————————————————-

[Karli Mather]

Are you seeing this performance issue in all your AX environments?  (DEV, TEST, UAT, PROD)

I believe the issues are primarily with production, however, I was just granted access to a new box for my personal development and the RAM sat around 50-60% when on AX for about 10 minutes doing non development tasks.The new box has  8GB  of RAM.

Did this issue just started recently? Have you tried using the AX trace tool to see what is causing the bad performance?

I’m new to the company and the company has only been on AX for about 6 months but it it is my understanding that performance has always been an issue or has been an issue for the majority of the time. I’m not sure if they’ve done tracing, I’ll check.

Do you have custom AX roles – is performance better for sys admin vs a user with custom AX role?

Pretty sure there are no custom roles but I do know there isn’t any noticeable difference between sys admin and other roles.

Were there lots of code change done recently?  If so, try running AXUtil Optimize.  More information can be found here – How to: Optimize a Model Store for Runtime [AX 2012]

I think there were a fair amount of customizations done to get it up and running, I will recommend this, thanks.

For the load balancing – are you using a dedicated AOS load balancer?  If not, do you have both the AOS listed in the client’s AX client config?

No dedicated AOS I will check on the configs. 

Thank you so much for your help!

——————————
Karli Mather
Fidelitone
Wauconda IL
————————————————————————-

[Jackie Olson]

I would ask for more RAM on your DEV box. 8GB is the bare bone requirement unless all you have on your DEV box is just the AOS and AX client. (no SQL, VS etc).  Can you ask for 16GB?

——————————
Jackie Olson
Stoneridge Software Inc.
Barnesville MN
————————————————————————-

[Scott Morley]

Karli,

A few more things to consider. Are you using the thick client on all the user’s desktops or are you running terminal services (with Citrix or Microsoft TS)? For the load balancing, is it being handled by the AOS servers (set in a cluster) or is it being handled by a network load balancer external to AX? If you can replicate the slowdown in a DEV or TEST environment then running AX tracing will be the best way to see what is going on. If it is really only happening in production and you don’t want to run a trace there (which is understandable) then you may need to install something like the DynamicsPerf tool to help capture where the slowdowns are.

If you have not used DynamicsPerf before, having someone who knows SQL Server will be really helpful. The download and guide can be found at DynamicsPerf 2.0 IS HERE !! – Dynamics AX in the Field

Microsoft		remove preview
DynamicsPerf 2.0 IS HERE !! – Dynamics AX in the Field View this on Microsoft >

——————————
Scott Morley
Principal Application Architect
OneNeck IT Solutions
Bend OR
————————————————————————-