From mboxrd@z Thu Jan 1 00:00:00 1970 From: Moshe Yudkowsky Subject: Documentation? failure to update-initramfs causes Infinite md loop on boot Date: Wed, 30 Jan 2008 16:32:46 -0600 Message-ID: <47A0FB0E.5020202@pobox.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Sender: linux-raid-owner@vger.kernel.org To: linux-raid@vger.kernel.org List-Id: linux-raid.ids I reformatted the disks in preparation to my move to a RAID1/RAID5 combination. I couldn't --stop the array (that should have told me something), so I removed ARRAY from mdadm.conf and restarted. I ran fdisk to create the proper partitions, and then I removed the /dev/md* and /dev/md/* entries in anticipation of creating the new ones. I then rebooted to pick up the new partitions I'd created. Now I can no longer boot, with this series of messages: md: md_import_device returned: -22 md: mdadm failed to add /dev/sdb2 to /dev/md/all: invalid argument mdadm: failed to RUN_ARRAY /dev/md/all: invalid argument md: sdc2 has invalid sb, not importing! Thousands of these go past, and there's no escape. That's quite a severe error. I'm going to boot on a rescue disk to fix this -- there's no other way I can think of to get out of this mess -- but I wonder if there ought to be documentation on the interaction between mdadm and update-initramfs. -- Moshe Yudkowsky * moshe@pobox.com * www.pobox.com/~moshe "Becoming the biggest banana republic in the world -- and without the bananas, at that -- is an unenviable prospect." -- Sergei Stepashin, Prime Minister of Russia