If a shared disk is reserved by one node, the other node may not boot up [315015]. During reboot, when the VxVM is coming up, the VM tries to read the private region data for all disks it can see, including the one reserved by the DiskReservation on some other node. The read goes through Dynamic Multipathing, which propagates only the EIO. Consequently, the rest of VM does not realize that a RESERVATION_CONFLICT has occurred, and retries the private region read. The rebooted node comes up only after the VM has done all the retries.

Workaround: Use vxdiskadm utility to remove all disks from VM that the DiskReservation uses. Reboot each node on which you remove the disks.