We had a lot of help from Sisense Support to migrate basically our entire setup; so not just cube + dashboards but also users, groups, data security, hierarchies etc. It is possible to do most of that via the REST APIs but I did investigate how that would work and it is really not pretty.
But if it's just Elasticubes and dashboards that need to be migrated, it should be possible to export the .smodel file for the cube(s) from the Windows web interface and import the cube into the Linux environment. You can then rebuild the cube and it should be fine.
Similarly for the dashboards -- export from Windows (can even be done in bulk) and re-import in Linux. If the data source is a cube with the same name on both ends, they should even be associated correctly already.
I *would* check those dashboards and cube -- After migration I noticed a few small and generally very subtle differences in the way Sisense Linux handles data import and queries to the cube as compared to Sisense Windows, so you may need to make some adjustments afterwards.
The main one on a cube level that comes to mind is importing arrays: on Windows, these were forced to text strings with a certain format. In earlier Linux versions the same thing happened but the format of the string was different. In current Linux versions, it appears that arrays are not imported at all from a database and you'll have to convert the array to string yourself in the import view.
Good Luck!
