It's built on top of GemStone/S which is an incredibly mature object database/persistance mechanism, while I can not speak for MagLev itself, if the guys at GemStone Inc, now VMware, are prepared to push this out as a v1.0, then I would be willing to bet yes.

Caveat is that while the Ruby implementation on top may not be 100% compatible with standard ruby (again I can't comment) the underlying technology is bulletproof, these guys have been doing object persistance w/ very high tps for 25+ years

This is an exciting technology, and if I worked with ruby at all, I would definitely be checking it out.