Max # of columns in source table

[mattmca]

I am working on importing a flat file (CSV) with a large number of columns (~350) into an elasticube. When I go to build the cube, even when only using a small number of sample rows, I get an error that says "Memory allocation has failed for <column name>. Please verify you have sufficient RAM and try to rebuild." We haven't had any memory issues on the server before, so not sure if this is a hardware limitation or a limitation of a maximum number of allowable columns. Has anybody else run into this issue before?

[KatieG]

Hi @mattmca - before getting into troubleshooting, please check if your Sisense instance is running on Windows or Linux (how to check here).

If it is Sisense on Linux - check the data groups setting for the default data group - by default the Reserved RAM is the minimum amount of RAM that must be free in order to build the cube. And the Max RAM is the max amount of RAM that the build process can request for building.

Its possible that a table that wide requires more RAM to process the build initially. For a test - you could make both values -1 so there is no limit - or make the Max RAM some acceptable value (under the total RAM on the machine).

Hope that helps!

[mattmca]

Hi @KatieG ,

I ended up getting the cube to build without an error when I ran it later, so I'm not sure what the original issue was. There may have been another load on the server that was competing for RAM when I was first running the build.

We are running Sisense on a Windows server. Do you know if there's a similar method for adjusting the reserved/max RAM for the build process on Sisense for Windows?

Thanks,

Matt

[KatieG]

Hi @mattmca - glad you were able to get it built! 

There is not a similar method in Windows to cap/adjust RAM consumption in the build process. This feature was developed for Sisense on Linux and was one of the unique advantages of developing Sisense on Linux + use of Kubernetes.