Will clusterware start if I lose just 1 voting disk out of the 3 I have (ASM normal redundancy)?

Hi,

I have a 2 nodes Linux 64bits RAC with 11.2.0.3 GI and database software.

I have my 3 voting disks in an ASM diskgroup in normal redundancy.

If I lose just 1 VD's disk, then clusterware remains up and running because 2 more VD's are accessible (majority rule). I have tested this. OK, that's what it is supposed to.

My question is:

If, after losing that VD, I reboot the clusterware (the missing VD is still inaccessible), will the clusterware start or will it stop starting waiting for that missing VD disk?

Which is the expected behavior?

I supposed clusterware would reboot complaining 1 VD's disk was missing (but it would restart eventually), but, in my tests it didn't. It stops starting because of that VD missing. Is that correct?