Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 2739
  • Last Modified:

XenApp / XenDesktop Stress Testing

Can I ask (please keep answers low tech) management freindly when you are migrating to a new thin client environment (3000 users!) what types of testing should be done to ensure the thin client infrastructure can perform adequately when all users are logged in and using the system,

I just see it must be difficult to simulate the same amount of usage unless you ask all 3000 of your users to access the test system, which isnt really practical.

What sort of stress testing do you perform and how on test environments to see if they are going to be able to do the job? Or what sort of watermarks/thresholds are you looking for that this is not going to take the strain?
4 Solutions
pma111Author Commented:
Not sure if stress testing is the correct terminology so would be interested to know what its called...
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
We start off slow, and build up the numbers, 5,10,20,50,100

but we also use scripts to load the servers, for proof of concept.

Citrix do have load testers, but we use Tevron's Apps

pma111Author Commented:
What exact tests are they (tevron/citrix load testers) running, and on the flip side, what watermarks are you looking for for potential "oh dear this is really struggling"?

What kind of things are the scripts actually "doing".

Is load testing the correct terminology then?
What does it mean to be "Always On"?

Is your cloud always on? With an Always On cloud you won't have to worry about downtime for maintenance or software application code updates, ensuring that your bottom line isn't affected.

pma111Author Commented:
And if say when you do your load testing the infrastructure is struggling, does that mean more investment, or how did you "tweak" things so it could handle the strain?
pma111Author Commented:
Or you may wish to answer in line with, if you were an auditor/risk professional, what evidence from the load/stress testing would you want for assurance that when users start utilising this infrastructure its going to perform and be fit for purpose. Based on both testing scope, findings/results, remidial action....
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
More Investment usually, or reduce loads, sometimes it could be re-tweaking or configuration, hence why it's important to stage, phase deployments to large numbers.

Because it's very difficult to model users concurrent behavior.

So testing/load testing simulation, is for worst case scenarios

e.g Monday Morning 9.00am, 3,000 users logons, can the infrastructure cope and deliver, and is logon time acceptable , with 60 seconds loading a desktop.

it would fail if it took 20 minutes to load a desktop!

So there needs to be a benchmark, users perceptions, expectations, etc
Automated testing has some advantages, but there is no substitution for real users.  If you can get a small group (maybe 10%?) of pilot users to really test either XenApp or XenDesktop and then extrapolate the results, that will give you some fairly good numbers.

So far as the test results, look for not only CPU and memory, but take a good look at disk IO and storage.  Disk IO is the real sleeper that can bite you later, and many people underestimate the impact of disk IO.

In particular, if you have several monitoring agents, your disk IO will creep up and make the system appear sluggish.  

Login VSI is pretty much the standard for automated testing, and a free version is available from LoginConsultants.com.
Ayman BakrSenior ConsultantCommented:
Performance or load testing would rather be more appropriate.

I would agree that there is no substitute for real users, however automation will give you a good indication what to have and from where to start. That is to say, with automation like EdgeSight for load testing you can determine how many sessions a server can tolerate before breaking down, how much resources are being used; then you can have a pilot environment with pilot users (representatives of your overall users in terms of application usage) as explained by Joharder at least 10% of your overall users.

If you want to do it manually, you can monitor the load on your servers using these, non-exhaustive yet most important, performance counters (besides the Processor and Memory counters):

Paging File % usage: High numbers for a long time indicate that you have too little RAM

Physical Disk - Avg Disk Queue Length: I believe usually should be less than 2 (a higher number is a clear indication of disk congestion)

Physical Disk - Avg Disk Sec/Read: Avg should be around 20 ms with spikes no higher than 50 ms. A problem in this is an indication of congestion in reading data from SAN

Physical Disk - Avg Dsik Sec/Write: same as above

System - Processor Queue Length: a number higher than 3 usually mean that processor is not sufficient or processors are very overloaded

Network Interface - Bytes Total/Sec: should not exceed 70% to the total bandwidth of the interface

Featured Post

Receive 1:1 tech help

Solve your biggest tech problems alongside global tech experts with 1:1 help.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now