On Tue, 7 Dec 2010 16:10:48 +0000 "Labun, Marcin" [off-list ref]
wrote:

quoted

From ee52735e3576dc998a837229ac5b9fb3ed1faeaf Mon Sep 17 00:00:00 2001

From: Marcin Labun <redacted>
Date: Wed, 30 Nov 2010 15:47:19 +0100
Subject: [PATCH 2/2] IMSM: do not rebuild the array if a non-redundant sub-array with failed disks is present

Now Intel metadata handler rebuilds all sub-arrays even if one of them
is non-redundant. In case of failed sub-array, failed disks are just replaced
with new ones in the metadata mapping. The data for failed disk is not restored
even the disk is present in the system. This fix requests to remove the failed
disk from container to let the process of rebuilding the array with failed
member. If the disk is physically pulled out of the system, the disk is removed
from container automatically by exiting udev rules.

This mostly makes sense, though...

 
+	/* 
+	 * If there are any failed disks check state of the other volume. 
+	 * Block rebuild if the other one is failed until failed disks
+	 * are removed from container.
+	 */

This comment was a lot clearer to me that the description at the top :-)

However:

+	if (failed) {
+		dprintf("found failed disks in %s, check if there is another"
+			"sub-array\n",
+			dev->volume);
+		/* check the state of the other volume allows for rebuild */
+		allowed = imsm_rebuild_allowed(a, (inst == 0) ? 1 : 0, failed);

                                                   ^^^^^^^^^^^^^^^^^^

This seems to imply that there are only ever at most 2 volumes in a
container.  Is that really true?  The rest of the code seems to assume that
there could be several.
If there can be more than two, then you need a loop over all the 'other'
devices to check that they are allowed to rebuild.

Thanks,
NeilBrown