Qlik Community

QlikView Administration

Discussion Board for collaboration on QlikView Management.

Announcements
Join “Do More With Qlik” with Mike Tarallo to get a first-hand look at the New Authoring Experience, June 29th: REGISTER HERE
cancel
Showing results for 
Search instead for 
Did you mean: 
f-lequeux
Contributor
Contributor

VM Qlikview 12 with AMD EPYC 7302P

Hello,

We had a qlikview E2620 server (2.1Ghz , 8 cores, 126 GB RAM) virtualized on proxmox. We decided to change the processor to bring more power.
We switched to an AMD EPYC 7302P (3Ghz, 16 cores, 256GB RAM).

We had during the transfer of the vm, quite a few problems with a windows process "system". It was consuming all the cpu resources at startup. It was doing a check on the ram, and all the processor was used. (99%).

My colleagues managed to solve this problem by dedicating 10% of the total ram to the hypervisor. Windows now starts normally.

I have some reloads that now take longer than before.

A small reload of 2 minutes now takes 4 minutes.
Some big reloads of one hour still take the same time while others take 20 minutes more.
Not all reloads are impacted by this change. I don't have any common point between reloads like for example the database or the use of a qvd.
Everything is random.
We have applied all the settings recommended by qlikview found here :
https://community.qlik.com/t5/Scalability/Quick-tips-8-Server-settings-for-best-performance/td-p/148...

What are we missing?
I was really expecting an improvement of reloads and performances.
Maybe not an improvement of RAW reloads, but at least an improvement of transformed qvd reloads.

Thanks to all.

Labels (1)
  • Other

3 Replies
David_Friend
Support
Support

Can you pick out one specific task and compare its Script and TaskLog to before you upgraded the server, see where the delta is?

korsikov
Partner
Partner

faced with strange behaviour after swich to AMD Epyc 7502p.

In Qlik Sense it's Fiber loop Error, but AMD server more powerful than previous intel one.

In QliKView - reload tasks crash with general script error (Rest connector reasone 999) 

And in general, the server is working slower

Albert_Candelario

Hello @f-lequeux !

I would also suggest @David_Friend 's approach to compare where the extra time is spent. For example you could use Notepad++ and use the plugin named Compare and see where the extra time is used to narrow down the investigation. Plus, were you using the same QlikView version during such time and only the CPU was changed?

Cheers!

Please, remember to mark the thread as solved once getting the correct answer