BookmarkSubscribeRSS Feed
☑ This topic is solved. Need further help from the community? Please sign in and ask a new question.
Achraf_Ougdal
Obsidian | Level 7

Hello Everyone,

 

I have a general question, We use SAS VA (Non-Distributed) on SAS 9.4 on Windows and we're facing some performance Issues (Server Crashing, Slow Reports etc...). We have some pretty huge data (A single star schema stored as a view has approx 230GB)

 

I am Wondering what do you use for your Servers ? I mean what are the Server Specs that you have.

Here's ours :

64Gb of RAM

x64 Intel Xeon E-2670, 2.6 Ghz

32 vCPU

500GB Disk Space

 

What do you think ?

1 ACCEPTED SOLUTION

Accepted Solutions
SASKiwi
PROC Star

From what I've heard VA star schemas are very respource-intensive and should be avoided if single analytics tables can do the job - this is the path we have chosen. Also we have had poor performance on some VA reports in the past. We engaged with SAS Tech Support to identify the cause of the problems and it turned out to be Java memory settings. I'd suggest you do the same so you can isolate the cause of your problems.

 

I agree 64GB memory sounds way too little especially for a non-distributed install. Our VA web-server alone has 40GB memory and occasionally runs short of memory. Check in VA Admininstrator to confirm what percentage of this memory is being consumed by your LASR tables.

 

Also if you are using VA 7.5, Environment Manager will tell you a lot about how your install is performing.

View solution in original post

4 REPLIES 4
Sam_SAS
SAS Employee

Hello,

 

I can't offer any insight into hardware requirements, but you might want to check whether the character columns in your data are wider than necessary. I had a table where the columns were all imported at 255 width, and reducing that to the actual width of the data values reduced the size of the data by more than 75%. 

 

Sam

 

 

Nigel_Pain
Lapis Lazuli | Level 10

I'd be inclined to have more RAM than that. The system requirements recommend a minimum of 16Gb per core (that would be 512Gb). We only have 8 cores, compared to your 32, but we have 128Gb RAM. That's probably over-egging it, to be honest, but that's mainly because VA doesn't get as much use as we expected when our servers were spec'd. Built as Windows Server 2012 R2 but we've recently upgraded to 2019.

SASKiwi
PROC Star

From what I've heard VA star schemas are very respource-intensive and should be avoided if single analytics tables can do the job - this is the path we have chosen. Also we have had poor performance on some VA reports in the past. We engaged with SAS Tech Support to identify the cause of the problems and it turned out to be Java memory settings. I'd suggest you do the same so you can isolate the cause of your problems.

 

I agree 64GB memory sounds way too little especially for a non-distributed install. Our VA web-server alone has 40GB memory and occasionally runs short of memory. Check in VA Admininstrator to confirm what percentage of this memory is being consumed by your LASR tables.

 

Also if you are using VA 7.5, Environment Manager will tell you a lot about how your install is performing.

SAS Innovate 2025: Call for Content

Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 25. Read more here about why you should contribute and what is in it for you!

Submit your idea!

Tips for filtering data sources in SAS Visual Analytics

See how to use one filter for multiple data sources by mapping your data from SAS’ Alexandria McCall.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 4 replies
  • 809 views
  • 4 likes
  • 5 in conversation