On Mon, 27 Dec 2010 19:24:47 -0800 John Jason Jordan <[email protected]> dijo:
>>Well, mdadm --query fails because I didn't specify a device. Of >>course, that is the whole problem. > >The correct command is "cat /proc/mdstat," since the arrays are >maintained in mdstat, which is just a text file. > >Turns out my arrays are named md0 to md3, and the query command goes >"mdadm --query /dev/md0" etc., although --detail gives better results. >And it appears that md1 is the bad one. > >The md1 array is composed of /dev/sda1 and /dev/sdb1, but the --detail >command says that /dev/sda1 has been removed. > >I'm trying to get my head around why removing /dev/sda1 would give me >the symptoms I am having. The only problem is that each dist-upgrade >installs a new kernel option in the Grub menu, but I cannot boot to >anything but the last Hardy kernel. Selecting any of the newer kernels >gives me a Grub error -15. So what did these dist-upgrades do, write a >kernel to a non-existent disk? It's trivially easy once you figure out how. All had to do was: mdadm /dev/md1 -a /dev/sda1 # the -a means "add" It took half an hour to rebuild the array. When I rebooted I could boot into the new kernels without a Grub error -15. I still don't understand exactly why. But it's fixed. I think. And it's currently dist-upgrading to Lucid. Later I'll disassemble the arrays. But that's an exercise for another day. _______________________________________________ PLUG mailing list [email protected] http://lists.pdxlinux.org/mailman/listinfo/plug
