I'm trying to test a setup of multiple EBS drive in a RAID configuration. What I'm looking for is what will happen, how will the raid (mdadm) handle a failure of one of the drives.

My setup is Ubuntu linux(official precise AMI), using mdadm for a 2-rive raid1 config.
I simulate a failure by doing a 'force-detach' of one of the drives through the API.

What happens on Ubuntu, CentOS and Amazon Linux AMIs is that the mount raid freezes/hangs, calls to variuous mdadm diagnostic functions hangs and a reboot is required.

RedHat AMIs are the only system where I can complete this tests successfully, the disk is marked as faulty and raid and the rest of the system continues its operation.

Whats special about the RedHat AMIs? Where do I begin to look? Is it kernel related?

I have checked logs for clues but I have nothing interesting to show.