Skip to main content

Has anyone noticed a performance degredation when running identical systems on identical hardware with the only difference being Windows server 2012 or 2016?

This is in a remote desktop environment but just one person logged in running 1 (9.2.0) Acucobol-GT runtime.

2016 is slower, in some cases when on a virtual platform by 50%.

I realise the "answer" will be to upgrade to the latest Acu version, but are there any known workarounds to speed it up?

Just had the scenario where an old VMware server was replaced with a shiny new one, bags of RAM, better quad core Xeon CPU, and the same system is running slower than the previous one. Must be a 2016/Acu interoperability thing as the old one on 2012 is faster than the new one :(

Has anyone noticed a performance degredation when running identical systems on identical hardware with the only difference being Windows server 2012 or 2016?

This is in a remote desktop environment but just one person logged in running 1 (9.2.0) Acucobol-GT runtime.

2016 is slower, in some cases when on a virtual platform by 50%.

I realise the "answer" will be to upgrade to the latest Acu version, but are there any known workarounds to speed it up?

Just had the scenario where an old VMware server was replaced with a shiny new one, bags of RAM, better quad core Xeon CPU, and the same system is running slower than the previous one. Must be a 2016/Acu interoperability thing as the old one on 2012 is faster than the new one :(

I'm not a virtualization expert but there may be something in the VM setup that is limiting resources. Even though the VM server has tons of resources a specific VM may not be allowing access to enough of it.

You should get at least an evaluation license for 10.2.1 and try it.  If it is also slow then that would indicate something in that environment.  It would be a good idea to test your application with version 9 on a different Windows Server 2016.

Also I'm wondering what is slow - screen display/navigation/update, file processing, something else?

If it is file processing where do the files live and how are they accessed?  Even if the files are local on that same server, if the directory where they reside is shared then Windows will implement the network protocols which slow things down.  Unsharing that directory could speed up file access.

 


I'm not a virtualization expert but there may be something in the VM setup that is limiting resources. Even though the VM server has tons of resources a specific VM may not be allowing access to enough of it.

You should get at least an evaluation license for 10.2.1 and try it.  If it is also slow then that would indicate something in that environment.  It would be a good idea to test your application with version 9 on a different Windows Server 2016.

Also I'm wondering what is slow - screen display/navigation/update, file processing, something else?

If it is file processing where do the files live and how are they accessed?  Even if the files are local on that same server, if the directory where they reside is shared then Windows will implement the network protocols which slow things down.  Unsharing that directory could speed up file access.

 

Thanks for the suggestions.

The files are normal flat vision files all stored locally in an unremarkable, unshared directory.

The slowdown is most apparent when running a large report, and the CPU and RAM are almost idle, but the hard disk activity is red with wrun32.exe taking 99% of the disk access, so it's apparent it's the disk access that is the bottleneck.

I guess I could play with the settings on the VM hard disk interface, but the fact remains that in an identical environment, 2012 is faster than 2016


Thanks for the suggestions.

The files are normal flat vision files all stored locally in an unremarkable, unshared directory.

The slowdown is most apparent when running a large report, and the CPU and RAM are almost idle, but the hard disk activity is red with wrun32.exe taking 99% of the disk access, so it's apparent it's the disk access that is the bottleneck.

I guess I could play with the settings on the VM hard disk interface, but the fact remains that in an identical environment, 2012 is faster than 2016

Hi

Did you find a solution i have the same problem!!!

Thank you

 


Hi

Did you find a solution i have the same problem!!!

Thank you

 

Unfortunately nothing solid Michel.

We still notice performance impact between 2012 and 2016. On an extremely well specified machine we get slightly better performance on a brand new 2016 VM than an old real 2012 server but nowhere near as much as 5 years of technological advancements (eg SSDs) should provide.

We also believe that  a real machine is better than a VM, and that VMware is slightly better than HyperV as a hypervisor.

We did try tweaking relevant cblconfig variables but nothing that made any significant difference.

Let me know what you learn as you experiment with this. Get a cushion to put on your desk to protect your head.


Has anyone noticed a performance degredation when running identical systems on identical hardware with the only difference being Windows server 2012 or 2016?

This is in a remote desktop environment but just one person logged in running 1 (9.2.0) Acucobol-GT runtime.

2016 is slower, in some cases when on a virtual platform by 50%.

I realise the "answer" will be to upgrade to the latest Acu version, but are there any known workarounds to speed it up?

Just had the scenario where an old VMware server was replaced with a shiny new one, bags of RAM, better quad core Xeon CPU, and the same system is running slower than the previous one. Must be a 2016/Acu interoperability thing as the old one on 2012 is faster than the new one :(

Is it possible that Windows Defender - or Endpoint Protection is helping cause the bottleneck. I believe you can designate exceptions for that virus software.


Unfortunately nothing solid Michel.

We still notice performance impact between 2012 and 2016. On an extremely well specified machine we get slightly better performance on a brand new 2016 VM than an old real 2012 server but nowhere near as much as 5 years of technological advancements (eg SSDs) should provide.

We also believe that  a real machine is better than a VM, and that VMware is slightly better than HyperV as a hypervisor.

We did try tweaking relevant cblconfig variables but nothing that made any significant difference.

Let me know what you learn as you experiment with this. Get a cushion to put on your desk to protect your head.

Hi

Sorry to ear that we are using 10.2.1 on a server 2016 and still the same problem.

We to have a customer that his using VM server 2016 and it is correct with the speed!

 


Has anyone noticed a performance degredation when running identical systems on identical hardware with the only difference being Windows server 2012 or 2016?

This is in a remote desktop environment but just one person logged in running 1 (9.2.0) Acucobol-GT runtime.

2016 is slower, in some cases when on a virtual platform by 50%.

I realise the "answer" will be to upgrade to the latest Acu version, but are there any known workarounds to speed it up?

Just had the scenario where an old VMware server was replaced with a shiny new one, bags of RAM, better quad core Xeon CPU, and the same system is running slower than the previous one. Must be a 2016/Acu interoperability thing as the old one on 2012 is faster than the new one :(

There was another forum discussion similar to this one:

https://community.microfocus.com/t5/Extend-and-AcuCOBOL-Forum/ACU-Slowness-with-W-Server-2012/m-p/1728171#M10902

setting registry key HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\TSFairShare\\Disk

EnableFairShare to 0

Look at this:

social.technet.microsoft.com/.../rd-session-host-slows-down-program