Forum Discussion

mjayaraman's avatar
mjayaraman
Cloud Apps
01-30-2023

Question regarding multi node build

In the linux server, if we use multi servers, we want to build the cubes in one server and copy to the other serving nodes and attach the cubes. 

We were doing it by copying the data folder in the Windows environment previously. Now we are upgrading it to linux, in the linux environment as well, do we have the cube data in a specific folder to copy from one server to another and attach it?

From the CLI commands help documents, i see a way to export the cube data into a file and then copy to another server, then import it in the destination server. In this approach, we see that it is unnecessary to export the data into file and it may take unnecessary storage or cpu time if we do it 15 to 20 times a day per cube.

Please advice on this. 

Data Models

2 Replies

Replies have been turned off for this discussion
  • irismaessen's avatar
    irismaessen
    Data Pipeline
    01-31-2023

    We used to have a similar setup when we were running Sisense on Windows with multiple servers -- copying the cube folders to a different server, attaching the cubes on the other one, deleting old cubes... we had a lot of scripting around it.

    One really *nice* thing about Sisense Linux in a multinode setup is... it takes care of these things. You can designate one or more nodes as build nodes, and the rest as query nodes and Sisense will arrange the rest. 

    It took quite a bit more effort to get it installed properly in our self-hosted case but we had help from Support and did some experimentation to get our disks and storage setup properly.