Has anyone else seen problems using linux software raid on large (2TB) disks with guid partition tables? We have a CentOS box that loses it during boot, and requires the raid be reassembled. Unfortunately, it's the only box I have with a bunch of 2TB disks, and it needs to go into production.
I would prefer to have a partition table on the disk, to ensure that the block devices are identical in size, especially if drives need to be replaced. Am I being too cautious, and I should simply build the array with the raw devices?