All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
To: dm-devel@redhat.com, christophe.varoqui@gmail.com
Subject: [PATCH] multipath-tools: prevent unexpected swapping of underlying LUNs
Date: Fri, 17 Aug 2012 17:40:39 +0900	[thread overview]
Message-ID: <502E0387.5070808@ce.jp.nec.com> (raw)
In-Reply-To: <5007C043.4000006@ce.jp.nec.com>

When you want to rename a multipath device
but the new alias is already used by other multipath device,
multipath-tools mistakenly reload a table for the original multipath device
to the other multipath device.
That could lead to very bad result, such as I/O error and data corruption.

This patch checks such a condition and gives up renaming with error log.

For example, suppose you have following 'bindings' file:
   
   # cat /etc/multipath/bindings 
   mpatha 212140084abcd0000
   mpathb 212150084abcd0000

and a logical volume 'VG/LV0' on top of mpathb,
which is on top of /dev/sde(8:64) and /dev/sdk(8:160):

   # dmsetup ls --tree
   mpatha (253:1)
    ├─ (8:144)
    └─ (8:48)
   VG-LV0 (253:2)
    └─mpathb (253:0)
       ├─ (8:160)
       └─ (8:64)

Then you decide to swap their names and change the 'bindings' as follows:

   # cat /etc/multipath/bindings 
   mpathb 212140084abcd0000
   mpatha 212150084abcd0000

you'll get this after 'service multipathd reload':

   # dmsetup ls --tree
   mpatha (253:1)
    ├─ (8:160)
    └─ (8:64)
   VG-LV0 (253:2)
    └─mpathb (253:0)
       ├─ (8:144)
       └─ (8:48)

Now you suddenly have 'VG/LV0' on top of /dev/sdd(8:48) and /dev/sdj(8:144),
that is obviously wrong and will corrupt data if you write to 'VG/LV0'.


diff -urpN multipath-tools.orig/libmultipath/configure.c multipath-tools.new/libmultipath/configure.c
--- multipath-tools.orig/libmultipath/configure.c	2012-07-19 15:17:40.368622358 +0900
+++ multipath-tools.new/libmultipath/configure.c	2012-08-17 17:11:07.691898774 +0900
@@ -150,12 +150,12 @@ static void
 select_action (struct multipath * mpp, vector curmp, int force_reload)
 {
 	struct multipath * cmpp;
+	struct multipath * cmpp_by_name;
 
-	cmpp = find_mp_by_alias(curmp, mpp->alias);
-
-	if (!cmpp) {
-		cmpp = find_mp_by_wwid(curmp, mpp->wwid);
+	cmpp = find_mp_by_wwid(curmp, mpp->wwid);
+	cmpp_by_name = find_mp_by_alias(curmp, mpp->alias);
 
+	if (!cmpp_by_name) {
 		if (cmpp) {
 			condlog(2, "%s: rename %s to %s", mpp->wwid,
 				cmpp->alias, mpp->alias);
@@ -169,17 +169,25 @@ select_action (struct multipath * mpp, v
 		return;
 	}
 
-	if (!find_mp_by_wwid(curmp, mpp->wwid)) {
-		condlog(2, "%s: remove (wwid changed)", cmpp->alias);
+	if (!cmpp) {
+		condlog(2, "%s: remove (wwid changed)", mpp->alias);
 		dm_flush_map(mpp->alias);
-		strncpy(cmpp->wwid, mpp->wwid, WWID_SIZE);
-		drop_multipath(curmp, cmpp->wwid, KEEP_PATHS);
+		strncpy(cmpp_by_name->wwid, mpp->wwid, WWID_SIZE);
+		drop_multipath(curmp, cmpp_by_name->wwid, KEEP_PATHS);
 		mpp->action = ACT_CREATE;
 		condlog(3, "%s: set ACT_CREATE (map wwid change)",
 			mpp->alias);
 		return;
 	}
 
+	if (cmpp != cmpp_by_name) {
+		condlog(2, "%s: unable to rename %s to %s (%s is used by %s)",
+			mpp->wwid, cmpp->alias, mpp->alias,
+			mpp->alias, cmpp_by_name->wwid);
+		mpp->action = ACT_NOTHING;
+		return;
+	}
+
 	if (pathcount(mpp, PATH_UP) == 0) {
 		mpp->action = ACT_NOTHING;
 		condlog(3, "%s: set ACT_NOTHING (no usable path)",

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

      reply	other threads:[~2012-08-17  8:40 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-07-19  8:07 [RFC PATCH] multipath-tools: Preventing silent swapping of underlying LUNs Jun'ichi Nomura
2012-08-17  8:40 ` Jun'ichi Nomura [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=502E0387.5070808@ce.jp.nec.com \
    --to=j-nomura@ce.jp.nec.com \
    --cc=christophe.varoqui@gmail.com \
    --cc=dm-devel@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.