COGNOiSe.com - The IBM Cognos Community

IBM Cognos Analytics Platform => Cognos Analytics => Reporting => Topic started by: cadams on 08 Mar 2018 11:25:27 PM

Title: Paramertised URL
Post by: cadams on 08 Mar 2018 11:25:27 PM
Hi,

As part of my upgrade from Cognos 10 to 11 I want to use Jmeter to compare a number of reports in both systems, and stress test the new environment. firstly I need to benchmark against C10.

My knowledge of jmeter and the ability to pass prompts is somewhat lacking, so I am trying to run reports via 'Report Views'. I'm using the storeID and can run reports with no problems, but when I use the storeID associated with a 'ReportView" it will not run.

Firstly what tools have you used to stress test Cognos environments.

Secondly is it actually possible to use a parameterized URL with a report view?
Title: Re: Paramertised URL
Post by: dougp on 09 Mar 2018 10:03:44 AM
Are you trying to stress test to see how much hardware to throw at Cognos?  That sounds like a lot of work  If you use the same hardware configuration for 11 as for 10, and your usage remains the same (same number of people, same frequency, same complexity -- basically, upgrading from 10 to 11), you should get about the same performance.

Are you trying to stress test with unrealistically high loads to see where Cognos fails?  I think that's IBM's job.
Title: Re: Paramertised URL
Post by: bdbits on 09 Mar 2018 11:35:24 AM
HP LoadRunner was used by my previous employer for this kind of thing. As I recall we mostly we used URLs to reports; you can pass report prompt values as part of the URL.

I've not used jmeter so I cannot offer you a comparison.
Title: Re: Paramertised URL
Post by: cadams on 11 Mar 2018 08:51:43 PM
Quote from: dougp on 09 Mar 2018 10:03:44 AM
Are you trying to stress test to see how much hardware to throw at Cognos?  That sounds like a lot of work  If you use the same hardware configuration for 11 as for 10, and your usage remains the same (same number of people, same frequency, same complexity -- basically, upgrading from 10 to 11), you should get about the same performance.

Are you trying to stress test with unrealistically high loads to see where Cognos fails?  I think that's IBM's job.

I wish it was that easy, I'm an external consultant that designed their 10 environment. 2 servers for load balancing and failover. The client has decided to change this for Cognos 11 and instead have a single server install. Mine is not to reason why... So I need to stress test as we're going from a load balanced service down to one, I'm going to have to do more configuration to ensure the same through put.