[PATCH 0 of 3] Patches to clean-up splitmirror code and fix bugs

[PATCH 0 of 3] Patches to clean-up splitmirror code and fix bugs

[Jonathan Brassow]

The following three patches are intended to fix bug 733114 - "mirror
conversion and mirror image splitting results in device i/o errors".

The first patch reverts a work-around that failed to address the issue
in the cluster environment.

The second addresses improper udev_flag settings for sub-LVs.

The third is a clean-up of the splitmirrors code.

 brassow

[PATCH 1 of 3] Revert initial solution to bug 733114

[Jonathan Brassow]

Revert initial solution to bug 733114 - I/O error message during splitmirror

The original commit comments can be located via this git commit ID:
	7d8e615c0b30fc2ef300c90378a51f01c328128c

There were three possible solutions to the original problem proposed in the
initial check-in.  The one chosen was as follows:
    2) Do like _remove_mirror_images does and suspend the original, then suspend
    the sub-lv (the error target), then resume the sub-lv, and finally resume the
    original LV.  This seems like extra pointless operations to me, but it doesn't
    produce the error message (although, I'm not sure why) and it allows us to
    leave the visible flag in place.
Turns out, the cluster also views the extra suspend/resume operations as
pointless too and ignores them.  So, this solution doesn't work in a cluster.
Further, I've noticed that in addition to the remote cluster nodes still getting
I/O errors from scanning the error target, they also have different LVM and
DM views of the same LV.  IOW, while the LVM level (gotten from the LVM metadata)
sees the correct name for the newly split LV, device-mapper still maintains the
old names.

Because the original fix failed to completely fix the problem (or work-around it)
and because a better solution must be found to address the additional cluster
issue of device renaming, I am reverting the above mentioned commit.

Index: LVM2/lib/metadata/mirror.c
===================================================================
--- LVM2.orig/lib/metadata/mirror.c
+++ LVM2/lib/metadata/mirror.c
@@ -673,10 +673,6 @@ static int _split_mirror_images(struct l
 		return 0;
 	}
 
-	/* Suspend temporary error target (see FIXME for resume below) */
-	if (sub_lv && !suspend_lv(sub_lv->vg->cmd, sub_lv))
-		return_0;
-
 	if (!vg_commit(mirrored_seg->lv->vg)) {
 		resume_lv(cmd, mirrored_seg->lv);
 		return 0;
@@ -685,42 +681,6 @@ static int _split_mirror_images(struct l
 	log_very_verbose("Updating \"%s\" in kernel", mirrored_seg->lv->name);
 
 	/*
-	 * FIXME:
-When an image is split from a 2-way mirror, the original mirror is converted to
-a linear device.  To do this, the top "layer" must be removed.  The segments
-are transferred from the sub-lv to the top-level LV and the link is severed. 
-The former sub-lv - having its segments transferred - now contains a temporary
-error target.
-
-When the original LV is resumed, the old sub-lv that now contains an error
-segment is activated and scanned.  This causes I/O error messages.  There are
-three ways to fix this problem:
-
-1) Do not set the sub-lv which contains the error target as "visible" before
-suspending the original LV.  This way, when the original is resumed, the sub-lv
-device node is not created and it is not scanned - avoiding the error messages.
- The problem with this approach is that if the machine crashes after the
-resume, it leaves the *hidden* LV in place and the user has a more difficult
-time noticing that it needs to be cleaned up.  Thus, this type of processing is
-frowned upon.
-
-2) Do like _remove_mirror_images does and suspend the original, then suspend
-the sub-lv (the error target), then resume the sub-lv, and finally resume the
-original LV.  This seems like extra pointless operations to me, but it does not
-produce the error message (although, I'm not sure why) and it allows us to
-leave the visible flag in place.  ** THIS IS THE CHOSEN SOLUTION HERE **
-
-3) Flag the sub-lv (error target) with a "do not scan" flag.  This seems like
-the cleanest approach, but I have been unable to find the method for doing
-this.  LVs get tagged in such a way by _get_udev_flags, but in this case the
-resume of the original LV also resumes the error target LV without running it
-through _get_udev_flags (likely because they are no longer linked).  Could
-there be something wrong in resume_lv?
-	*/
-	if (sub_lv && !resume_lv(sub_lv->vg->cmd, sub_lv))
-		return_0;
-
-	/*
 	 * Resume the mirror - this also activates the visible, independent
 	 *                     soon-to-be-split sub-LVs
 	 */

[PATCH 2 of 3] Fix udev flags on sub-lvs

[Jonathan Brassow]

This patch attempts to fix issues with improper udev flags on sub-LVs.

The current code does not always assign proper udev flags to sub-LVs (e.g.
mirror images and log LVs).  This shows up especially during a splitmirror
operation in which an image is split off from a mirror to form a new LV.

A mirror with a disk log is actually composed of 4 different LVs: the 2
mirror images, the log, and the top-level LV that "glues" them all together.
When a 2-way mirror is split into two linear LVs, two of those LVs must be
removed.  The segments of the image which is not split off to form the new
LV are transferred to the top-level LV.  This is done so that the original
LV can maintain its major/minor, UUID, and name.  The sub-lv from which the
segments were transferred gets an error segment as a transitory process
before it is eventually removed.  (Note that if the error target was not put
in place, a resume_lv would result in two LVs pointing to the same segment!
If the machine crashes before the eventual removal of the sub-LV, the result
would be a residual LV with the same mapping as the original (now linear) LV.)
So, the two LVs that need to be removed are now the log device and the sub-LV
with the error segment.  If udev_flags are not properly set, a resume will
cause the error LV to come up and be scanned by udev.  This causes I/O errors.
Additionally, when udev scans sub-LVs (or former sub-LVs), it can cause races
when we are trying to remove those LVs.  This is especially bad during failure
conditions.

When the mirror is suspended, the top-level along with it's sub-LVs are
suspended.  The changes (now 2 linear devices and the yet-to-be-removed log
and error LV) are committed.  When the resume takes place on the original
LV, there are no longer links to the other sub-lvs through the LVM metadata.
The links are implicitly handled by querying the kernel for a list of
dependencies.  This is done in the '_add_dev' function (which is recursively
called for each dependency found) - called through the following chain:
	_add_dev
	dm_tree_add_dev_with_udev_flags
	<*** DM / LVM divide ***>
	_add_dev_to_dtree
	_add_lv_to_dtree
	_create_partial_dtree
	_tree_action
	dev_manager_activate
	_lv_activate_lv
	_lv_resume
	lv_resume_if_active
When udev flags are calculated by '_get_udev_flags', it is done by referencing
the 'logical_volume' structure.  Those flags are then passed down into
'dm_tree_add_dev_with_udev_flags', which in turn passes them to '_add_dev'.
Unfortunately, when '_add_dev' is finding the dependencies, it has no way to
calculate their proper udev_flags.  This is because it is below the DM/LVM
divide - it doesn't have access to the logical_volume structure.  In fact,
'_add_dev' simply reuses the udev_flags given for the initial device!  This
virtually guarentees the udev_flags are wrong for all the dependencies unless
they are reset by some other mechanism.  The current code provides no such
mechanism.  Even if '_add_new_lv_to_dtree' were called on the sub-devices -
which it isn't - entries already in the tree are simply passed over, failing
to reset any udev_flags.  The solution must retain its implicit nature of
discovering dependencies and be able to go back over the dependencies found
to properly set the udev_flags.

There are probably multiple levels at which a solution could be done (either
up or down a function in the call chain from where I've placed it).  That
position may significantly alter the way in which the solution is coded, but
I don't think it will change the general idea.  I'm hoping that others have
some insight on where the best place to make the necessary changes is.

My solution simply calls a new function before leaving '_add_new_lv_to_dtree'
that iterates over the dtree nodes to properly reset the udev_flags of any
children.  It is important that this function occur after the '_add_dev' has
done its job of querying the kernel for a list of dependencies.  It is this
list of children that we use to look up their respective LVs and properly
calculate the udev_flags.

This solution has worked for single machine, cluster, and cluster w/ exclusive
activation.




Index: LVM2/libdm/libdm-deptree.c
===================================================================
--- LVM2.orig/libdm/libdm-deptree.c
+++ LVM2/libdm/libdm-deptree.c
@@ -720,6 +720,16 @@ struct dm_tree_node *dm_tree_add_new_dev
 	return node;
 }
 
+void dm_tree_node_set_udev_flags(struct dm_tree_node *dnode, uint16_t udev_flags)
+{
+	if (udev_flags == dnode->udev_flags)
+		log_debug("%s udev_flags already set to 0x%x",
+			  dnode->name, udev_flags);
+	else
+		log_debug("Reseting %s udev_flags (s/0x%x/0x%x)",
+			  dnode->name, dnode->udev_flags, udev_flags);
+	dnode->udev_flags = udev_flags;
+}
 
 void dm_tree_node_set_read_ahead(struct dm_tree_node *dnode,
 				 uint32_t read_ahead,
Index: LVM2/lib/activate/dev_manager.c
===================================================================
--- LVM2.orig/lib/activate/dev_manager.c
+++ LVM2/lib/activate/dev_manager.c
@@ -1524,6 +1524,41 @@ static int _add_segment_to_dtree(struct 
 	return 1;
 }
 
+static int set_udev_flags_for_children(struct dev_manager *dm,
+				       struct volume_group *vg,
+				       struct dm_tree_node *dnode)
+{
+	void *handle = NULL;
+	struct dm_tree_node *child;
+	const struct dm_info *info;
+	char *vgname, *lvname, *layer;
+	struct lv_list *lvl;
+
+	while ((child = dm_tree_next_child(&handle, dnode, 0)) &&
+	       (info  = dm_tree_node_get_info(child)) && info->exists) {
+		if (!dm_split_lvm_name(dm->mem, dm_tree_node_get_name(child),
+				       &vgname, &lvname, &layer)) {
+			log_error("Failed to split DM name, %s",
+				  dm_tree_node_get_name(child));
+			return 0;
+		}
+
+		if (!(lvl = find_lv_in_vg(vg, lvname))) {
+			log_debug("Failed to find %s in %s (due to rename?)",
+				  lvname, vg->name);
+			continue;
+		}
+
+		log_debug("Resetting udev flags for %s (a child of %s)",
+			  dm_tree_node_get_name(dnode),
+			  dm_tree_node_get_name(dnode));
+		dm_tree_node_set_udev_flags(child, _get_udev_flags(dm, lvl->lv,
+								   layer));
+	}
+
+	return 1;
+}
+
 static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 				struct logical_volume *lv, struct lv_activate_opts *laopts,
 				const char *layer)
@@ -1628,6 +1663,9 @@ static int _add_new_lv_to_dtree(struct d
 			if (!_add_new_lv_to_dtree(dm, dtree, sl->seg->lv, laopts, NULL))
 				return_0;
 
+	if (!set_udev_flags_for_children(dm, lv->vg, dnode))
+		return_0;
+
 	return 1;
 }
 
Index: LVM2/libdm/libdevmapper.h
===================================================================
--- LVM2.orig/libdm/libdevmapper.h
+++ LVM2/libdm/libdevmapper.h
@@ -535,6 +535,8 @@ int dm_tree_node_add_replicator_dev_targ
 					   uint32_t slog_region_size);
 /* End of Replicator API */
 
+void dm_tree_node_set_udev_flags(struct dm_tree_node *node, uint16_t udev_flags);
+
 void dm_tree_node_set_presuspend_node(struct dm_tree_node *node,
 				      struct dm_tree_node *presuspend_node);
 

[PATCH 3 of 3] Clean-up splitmirror code

[Jonathan Brassow]

Clean-up splitmirrors code.

I've attempted to clean-up the splitmirrors code to make it easier to
understand with fewer operations.  I've tried to reduce the number of
metadata operations without compromising the intermediate stages which
are necessary for easy clean-up in the even of failure.

These changes now correctly handle cluster situations - including exclusive
cluster mirrors.

Index: LVM2/lib/metadata/mirror.c
===================================================================
--- LVM2.orig/lib/metadata/mirror.c
+++ LVM2/lib/metadata/mirror.c
@@ -394,6 +394,19 @@ activate_lv:
 	return 0;
 }
 
+static int inkind_activate_lv(struct logical_volume *model,
+			      struct logical_volume *lv)
+{
+	if (lv_is_active_exclusive_locally(model)) {
+		if (!activate_lv_excl(lv->vg->cmd, lv))
+			return_0;
+	} else {
+		if (!activate_lv(lv->vg->cmd, lv))
+			return_0;
+	}
+	return 1;
+}
+
 /*
  * Delete independent/orphan LV, it must acquire lock.
  */
@@ -417,13 +430,9 @@ static int _delete_lv(struct logical_vol
 		}
 	}
 
-	if (lv_is_active_exclusive_locally(lv)) {
-		if (!activate_lv_excl(cmd, lv))
-			return_0;
-	} else {
-		if (!activate_lv(cmd, lv))
-			return_0;
-	}
+	// FIXME: shouldn't the activation type be based on mirror_lv, not lv?
+	if (!inkind_activate_lv(lv, lv))
+		return_0;
 
 	sync_local_dev_names(lv->vg->cmd);
 	if (!deactivate_lv(cmd, lv))
@@ -584,7 +593,7 @@ static int _split_mirror_images(struct l
 	struct logical_volume *detached_log_lv = NULL;
 	struct lv_segment *mirrored_seg = first_seg(lv);
 	struct dm_list split_images;
-	struct lv_list *lvl, *new_lvl = NULL;
+	struct lv_list *lvl;
 	struct cmd_context *cmd = lv->vg->cmd;
 
 	if (!(lv->status & MIRRORED)) {
@@ -626,12 +635,18 @@ static int _split_mirror_images(struct l
 		sub_lv = seg_lv(mirrored_seg, mirrored_seg->area_count);
 
 		sub_lv->status &= ~MIRROR_IMAGE;
-		lv_set_visible(sub_lv);
 		release_lv_segment_area(mirrored_seg, mirrored_seg->area_count,
 					mirrored_seg->area_len);
 
 		log_very_verbose("%s assigned to be split", sub_lv->name);
 
+		if (!new_lv) {
+			lv_set_visible(sub_lv);
+			new_lv = sub_lv;
+			continue;
+		}
+
+		/* If there is more than one image being split, add to list */
 		lvl = dm_pool_alloc(lv->vg->vgmem, sizeof(*lvl));
 		if (!lvl) {
 			log_error("lv_list alloc failed");
@@ -640,90 +655,7 @@ static int _split_mirror_images(struct l
 		lvl->lv = sub_lv;
 		dm_list_add(&split_images, &lvl->list);
 	}
-	sub_lv = NULL;
-
-	/*
-	 * If no more mirrors, remove mirror layer.
-	 * The sub_lv is removed entirely later - leaving
-	 * only the top-level (now linear) LV.
-	 */
-	if (mirrored_seg->area_count == 1) {
-		sub_lv = seg_lv(mirrored_seg, 0);
-		sub_lv->status &= ~MIRROR_IMAGE;
-		lv_set_visible(sub_lv);
-		detached_log_lv = detach_mirror_log(mirrored_seg);
-		if (!remove_layer_from_lv(lv, sub_lv))
-			return_0;
-		lv->status &= ~MIRRORED;
-		lv->status &= ~LV_NOTSYNCED;
-	}
 
-	if (!vg_write(mirrored_seg->lv->vg)) {
-		log_error("Intermediate VG metadata write failed.");
-		return 0;
-	}
-
-	/*
-	 * Suspend the mirror - this includes all the sub-LVs and
-	 *                      soon-to-be-split sub-LVs
-	 */
-	if (!suspend_lv(cmd, mirrored_seg->lv)) {
-		log_error("Failed to lock %s", mirrored_seg->lv->name);
-		vg_revert(mirrored_seg->lv->vg);
-		return 0;
-	}
-
-	if (!vg_commit(mirrored_seg->lv->vg)) {
-		resume_lv(cmd, mirrored_seg->lv);
-		return 0;
-	}
-
-	log_very_verbose("Updating \"%s\" in kernel", mirrored_seg->lv->name);
-
-	/*
-	 * Resume the mirror - this also activates the visible, independent
-	 *                     soon-to-be-split sub-LVs
-	 */
-	if (!resume_lv(cmd, mirrored_seg->lv)) {
-		log_error("Problem resuming %s", mirrored_seg->lv->name);
-		return 0;
-	}
-
-	/* Remove original mirror layer if it has been converted to linear */
-	if (sub_lv && !_delete_lv(lv, sub_lv))
-		return_0;
-
-	/* Remove the log if it has been converted to linear */
-	if (detached_log_lv && !_delete_lv(lv, detached_log_lv))
-		return_0;
-
-	/*
-	 * Step 2:
-	 *   The original mirror has been changed and we now have visible,
-	 *   independent, not-yet-renamed, active sub-LVs.  We must:
-	 *   - deactivate the split sub-LVs
-	 *   - rename them
-	 *   - form new mirror if necessary
-	 *   - commit VG changes
-	 *   - activate the new LV
-	 */
-	sync_local_dev_names(lv->vg->cmd);
-	dm_list_iterate_items(lvl, &split_images) {
-		if (!new_lv) {
-			/* Grab 1st sub-LV for later */
-			new_lvl = lvl;
-			new_lv = lvl->lv;
-		}
-
-		sub_lv = lvl->lv;
-		if (!deactivate_lv(cmd, sub_lv)) {
-			log_error("Failed to deactivate former mirror image, %s",
-				  sub_lv->name);
-			return 0;
-		}
-	}
-
-	dm_list_del(&new_lvl->list);
 	new_lv->name = dm_pool_strdup(lv->vg->vgmem, split_name);
 	if (!new_lv->name) {
 		log_error("Unable to rename newly split LV");
@@ -780,19 +712,76 @@ static int _split_mirror_images(struct l
 		init_mirror_in_sync(1);
 	}
 
-	if (!vg_write(mirrored_seg->lv->vg) ||
-	    !vg_commit(mirrored_seg->lv->vg))
-		return_0;
+	sub_lv = NULL;
+
+	/*
+	 * If no more mirrors, remove mirror layer.
+	 * The sub_lv is removed entirely later - leaving
+	 * only the top-level (now linear) LV.
+	 */
+	if (mirrored_seg->area_count == 1) {
+		sub_lv = seg_lv(mirrored_seg, 0);
+		sub_lv->status &= ~MIRROR_IMAGE;
+		lv_set_visible(sub_lv);
+		detached_log_lv = detach_mirror_log(mirrored_seg);
+		if (!remove_layer_from_lv(lv, sub_lv))
+			return_0;
+		lv->status &= ~MIRRORED;
+		lv->status &= ~LV_NOTSYNCED;
+	}
+
+	if (!vg_write(mirrored_seg->lv->vg)) {
+		log_error("Intermediate VG metadata write failed.");
+		return 0;
+	}
+
+	/*
+	 * Suspend the mirror - this includes all the sub-LVs and
+	 *                      soon-to-be-split sub-LVs
+	 */
+	if (!suspend_lv(cmd, mirrored_seg->lv)) {
+		log_error("Failed to lock %s", mirrored_seg->lv->name);
+		vg_revert(mirrored_seg->lv->vg);
+		return 0;
+	}
+
+	if (!vg_commit(mirrored_seg->lv->vg)) {
+		resume_lv(cmd, mirrored_seg->lv);
+		return 0;
+	}
+
+	log_very_verbose("Updating \"%s\" in kernel", mirrored_seg->lv->name);
+
+	/*
+	 * Resume the mirror - this also activates the visible, independent
+	 *                     soon-to-be-split sub-LVs
+	 */
+	if (!resume_lv(cmd, mirrored_seg->lv)) {
+		log_error("Problem resuming %s", mirrored_seg->lv->name);
+		return 0;
+	}
 
-	/* Bring newly split-off LV into existence */
-	if (!activate_lv(cmd, new_lv)) {
-		log_error("Failed to activate newly split LV, %s",
-			  new_lv->name);
+	/*
+	 * Recycle newly split LV so it is properly renamed.
+	 *   Cluster requires the extra deactivate/activate calls.
+	 */
+	if (vg_is_clustered(lv->vg) &&
+	    (!deactivate_lv(cmd, new_lv) || !inkind_activate_lv(lv, new_lv))) {
+		log_error("Failed to rename newly split LV in the kernel");
+		return 0;
+	}
+	if (!suspend_lv(cmd, new_lv) || !resume_lv(cmd, new_lv)) {
+		log_error("Failed to rename newly split LV in the kernel");
 		return 0;
 	}
 
-	log_very_verbose("%" PRIu32 " image(s) detached from %s",
-			 split_count, lv->name);
+	/* Remove original mirror layer if it has been converted to linear */
+	if (sub_lv && !_delete_lv(lv, sub_lv))
+		return_0;
+
+	/* Remove the log if it has been converted to linear */
+	if (detached_log_lv && !_delete_lv(lv, detached_log_lv))
+		return_0;
 
 	return 1;
 }

[PATCH 2 of 3] Fix udev flags on sub-lvs

[Jonathan Brassow]

Refresh patch to update the logging and ignore non-LVM devices when
resetting udev flags.

 brassow

This patch attempts to fix issues with improper udev flags on sub-LVs.

The current code does not always assign proper udev flags to sub-LVs (e.g.
mirror images and log LVs).  This shows up especially during a splitmirror
operation in which an image is split off from a mirror to form a new LV.

A mirror with a disk log is actually composed of 4 different LVs: the 2
mirror images, the log, and the top-level LV that "glues" them all together.
When a 2-way mirror is split into two linear LVs, two of those LVs must be
removed.  The segments of the image which is not split off to form the new
LV are transferred to the top-level LV.  This is done so that the original
LV can maintain its major/minor, UUID, and name.  The sub-lv from which the
segments were transferred gets an error segment as a transitory process
before it is eventually removed.  (Note that if the error target was not put
in place, a resume_lv would result in two LVs pointing to the same segment!
If the machine crashes before the eventual removal of the sub-LV, the result
would be a residual LV with the same mapping as the original (now linear) LV.)
So, the two LVs that need to be removed are now the log device and the sub-LV
with the error segment.  If udev_flags are not properly set, a resume will
cause the error LV to come up and be scanned by udev.  This causes I/O errors.
Additionally, when udev scans sub-LVs (or former sub-LVs), it can cause races
when we are trying to remove those LVs.  This is especially bad during failure
conditions.

When the mirror is suspended, the top-level along with it's sub-LVs are
suspended.  The changes (now 2 linear devices and the yet-to-be-removed log
and error LV) are committed.  When the resume takes place on the original
LV, there are no longer links to the other sub-lvs through the LVM metadata.
The links are implicitly handled by querying the kernel for a list of
dependencies.  This is done in the '_add_dev' function (which is recursively
called for each dependency found) - called through the following chain:
	_add_dev
	dm_tree_add_dev_with_udev_flags
	<*** DM / LVM divide ***>
	_add_dev_to_dtree
	_add_lv_to_dtree
	_create_partial_dtree
	_tree_action
	dev_manager_activate
	_lv_activate_lv
	_lv_resume
	lv_resume_if_active
When udev flags are calculated by '_get_udev_flags', it is done by referencing
the 'logical_volume' structure.  Those flags are then passed down into
'dm_tree_add_dev_with_udev_flags', which in turn passes them to '_add_dev'.
Unfortunately, when '_add_dev' is finding the dependencies, it has no way to
calculate their proper udev_flags.  This is because it is below the DM/LVM
divide - it doesn't have access to the logical_volume structure.  In fact,
'_add_dev' simply reuses the udev_flags given for the initial device!  This
virtually guarentees the udev_flags are wrong for all the dependencies unless
they are reset by some other mechanism.  The current code provides no such
mechanism.  Even if '_add_new_lv_to_dtree' were called on the sub-devices -
which it isn't - entries already in the tree are simply passed over, failing
to reset any udev_flags.  The solution must retain its implicit nature of
discovering dependencies and be able to go back over the dependencies found
to properly set the udev_flags.

There are probably multiple levels at which a solution could be done (either
up or down a function in the call chain from where I've placed it).  That
position may significantly alter the way in which the solution is coded, but
I don't think it will change the general idea.  I'm hoping that others have
some insight on where the best place to make the necessary changes is.

My solution simply calls a new function before leaving '_add_new_lv_to_dtree'
that iterates over the dtree nodes to properly reset the udev_flags of any
children.  It is important that this function occur after the '_add_dev' has
done its job of querying the kernel for a list of dependencies.  It is this
list of children that we use to look up their respective LVs and properly
calculate the udev_flags.

This solution has worked for single machine, cluster, and cluster w/ exclusive
activation.




Index: LVM2/libdm/libdm-deptree.c
===================================================================
--- LVM2.orig/libdm/libdm-deptree.c
+++ LVM2/libdm/libdm-deptree.c
@@ -720,6 +720,18 @@ struct dm_tree_node *dm_tree_add_new_dev
 	return node;
 }
 
+void dm_tree_node_set_udev_flags(struct dm_tree_node *dnode, uint16_t udev_flags)
+
+{
+	struct dm_info *dinfo = &dnode->info;
+
+	if (udev_flags != dnode->udev_flags)
+		log_debug("Resetting %s (%" PRIu32 ":%" PRIu32
+			  ") udev_flags (s/0x%X/0x%X)",
+			  dnode->name, dinfo->major, dinfo->minor,
+			  dnode->udev_flags, udev_flags);
+	dnode->udev_flags = udev_flags;
+}
 
 void dm_tree_node_set_read_ahead(struct dm_tree_node *dnode,
 				 uint32_t read_ahead,
Index: LVM2/lib/activate/dev_manager.c
===================================================================
--- LVM2.orig/lib/activate/dev_manager.c
+++ LVM2/lib/activate/dev_manager.c
@@ -1524,6 +1524,60 @@ static int _add_segment_to_dtree(struct 
 	return 1;
 }
 
+static int set_udev_flags_for_children(struct dev_manager *dm,
+				       struct volume_group *vg,
+				       struct dm_tree_node *dnode)
+{
+	void *handle = NULL;
+	struct dm_tree_node *child;
+	const char *uuid;
+	const struct dm_info *info;
+	char *vgname, *lvname, *layer;
+	struct lv_list *lvl;
+
+	while ((child = dm_tree_next_child(&handle, dnode, 0))) {
+		/* Ignore root node */
+		if (!(info  = dm_tree_node_get_info(child)) || !info->exists)
+			continue;
+
+		if (!(uuid = dm_tree_node_get_uuid(child))) {
+			log_error(INTERNAL_ERROR
+				  "Failed to get uuid for %" PRIu32 ":%" PRIu32,
+				  info->major, info->minor);
+			continue;
+		}
+
+		/* Ignore non-LVM devices */
+		if (strncmp(uuid, UUID_PREFIX, sizeof(UUID_PREFIX) - 1)) {
+			log_debug("Not resetting udev flags for non-LVM "
+				  "device %" PRIu32 ":%" PRIu32,
+				  info->major, info->minor);
+			continue;
+		}
+
+		if (!dm_split_lvm_name(dm->mem, dm_tree_node_get_name(child),
+				       &vgname, &lvname, &layer)) {
+			log_error("Failed to split DM name, %s",
+				  dm_tree_node_get_name(child));
+			return 0;
+		}
+
+		if (!(lvl = find_lv_in_vg(vg, lvname)))
+			/*
+			 * Rename of an LV will cause it not to be found, but
+			 * in these cases we should allow udev to rename the
+			 * device - not skip over it because it is a mirror
+			 * sub-lv or something similar
+			 */
+			continue;
+
+		dm_tree_node_set_udev_flags(child,
+					    _get_udev_flags(dm, lvl->lv, layer));
+	}
+
+	return 1;
+}
+
 static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 				struct logical_volume *lv, struct lv_activate_opts *laopts,
 				const char *layer)
@@ -1628,6 +1682,9 @@ static int _add_new_lv_to_dtree(struct d
 			if (!_add_new_lv_to_dtree(dm, dtree, sl->seg->lv, laopts, NULL))
 				return_0;
 
+	if (!set_udev_flags_for_children(dm, lv->vg, dnode))
+		return_0;
+
 	return 1;
 }
 
Index: LVM2/libdm/libdevmapper.h
===================================================================
--- LVM2.orig/libdm/libdevmapper.h
+++ LVM2/libdm/libdevmapper.h
@@ -535,6 +535,8 @@ int dm_tree_node_add_replicator_dev_targ
 					   uint32_t slog_region_size);
 /* End of Replicator API */
 
+void dm_tree_node_set_udev_flags(struct dm_tree_node *node, uint16_t udev_flags);
+
 void dm_tree_node_set_presuspend_node(struct dm_tree_node *node,
 				      struct dm_tree_node *presuspend_node);