Tension soil stress below underground water tanks (UGWT)

[NewbieInSE]

Dear Engineers,
In most of the cases when designing these UGWT, I found that, soil tension occurs in the following two cases:
1. Buoyancy load combination.
2. If the tank bottom slab is actually a mat, and the walls support two or three storied structure. In this case, the vertical load coming from above would alone produce tension in the soil at the middle of the span (SURROUNDED BY WALLS).

Is this tension (250 psf etc.) in soil negligible whatever be the source (buoyancy or vertical dead, live)?
Pls consider that, the empty tank is stable considering it's dead load, some soil load on the mat extension beyond the wall etc.

I'm attaching a figure just for visualization.

 

[EZBuilding]

This looks like a SAFE model. Have you created nonlinear loadcases which allows for uplift?

 

[NewbieInSE]

It's actually in ETABS model. I can make the buoyancy load combination nonlinear, ETABS doesn't have "allow uplift". However i made the area spring compression only, so if ETABS can solve the equation it won't show tension.
In that case, tension isn't shown at the middle of the slab for the particular load case / combo which have been made nonlinear.
I have several questions regarding this,
1. If tension occurs in the soil, do i have to make all the load cases/combos nonlinear so that structural members are designed for forces when the soil stress isn't tension? Or should i only make the combo nonlinear for which there is tension, in this case probably only the foundation stress check combo with buoyancy.