Qlik Community

Ask a Question

QlikView App Dev

Discussion Board for collaboration related to QlikView App Development.

Announcements
QlikView forum consolidation is complete. Labels are now required. LEARN ABOUT LABELS
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

Qlikview with RAM 700 GB, but still slow

Dear All,

I have 2 Documents QVW that one of them has size more than 100 GB (for example its called A) and the other is 16 GB (it's called B), .

A includes complex data models: there are 6 Facts and (more than) 20 Dimensions. I use link table to avoid synthetic keys. One of facts has a 1.5 billion rows. It contain complex chart, maybe there are 60-70 charts that is grouped in 3 pages.

B also includes complex data models: there are 3 Facts and 10 Dimensions. Two of facts has a 56 million rows. For UI, it has 7 pages. For each page, I only show straight table. But, in the straight table there are some calculation such as: multiplication, division, etc. And for some page, there are some specific conditions, so I trigger if user jump to page X, some filters are locked in page X.

In QMC --> System --> Setup --> QVS --> Performance, I have set 'Object Calculation Time Limit' until 5 minutes.

And In QMC --> Documents --> Reload, I have check 'Preload' box.

When I publish it to Qlikview Access Point:

For A: I can't open this dashboard from web. It always show message 'Lost Connection to Server', whereas the user license only dedicated to me.

For B: I have to wait 2-3 minutes until it open perfectly. When I select filter, there is also loading time (maybe 5-10 seconds) until it is filtered. Our user complain about this situations. I don't know what's happened with this.

Server is dedicated for Qlikview only and it has 750 GB RAM. But, the usage of RAM is very low. I don't know is there any configuration in Qlikview or in Server in order to maximize the usage of RAM?

Need your help to identify what's wrong with this situations.

Please suggestions

Thanks before,

Regards,

Yulindo

2 Replies
marcus_sommer
MVP & Luminary
MVP & Luminary

Beside some points on the environment (VM or physical machine, type of processors, various configurations) and if compressing is enabled (it needs some time to (de)compress an application - and if compression is enabled by a 100 GB application 750 GB RAM might not enough to handle it) or section access is used (these creates some overhead by opening an application) I think that your datamodel and your gui-design isn't really suitable for such a large application.

Large applications should be designed more in the direction of a star-scheme and the use of link-tables should be avoided. Further it should be tried to use simple calculations without nested if-loops, calculated dimensions, aggr-functions, dimension/expression-groups, containers and the use of triggers.

Further you should check to load only fields which are really used - Document Analyzer V3.0 Update Available - especially by high cardinality fields like record-id's or timestamps - The Importance Of Being Distinct.

Another good starting point could be: Advanced topics for creating a qlik datamodel

- Marcus

gregortvw
Contributor III
Contributor III

Hi,

I've a suggestion for the opening problem for A:

if you're using IIS:

Go to IIS Manager->Sites->QvAjaxZfc-Right Click->Explore->Web.config

edit the line

<httpRuntime requestValidationMode="2.0" /> to  <httpRuntime requestValidationMode="2.0" executionTimeout="900" />

Maybe in your partikular case, you'll need higher timeouts. (900 Seconds in my example)

On you second issue:

try to monitor CPU/RAM usage while using the app B, or look, if the HDD/SAN/Antivirus/other stuff is bottlenecking the connection between user and QVS

Also couple further questions:

How's your CPU configuration? (number/affinity)

Are there more apps?

Are the reloads scheduled? over night?

Regards,

Gregor