[prev in list] [next in list] [prev in thread] [next in thread] 

List:       lvm-devel
Subject:    [lvm-devel] LVM2 patch to clean-up --resync on cluster mirrors
From:       Jonathan Brassow <jbrassow () redhat ! com>
Date:       2008-05-15 19:45:26
Message-ID: 1210880726.12317.3.camel () hydrogen
[Download RAW message or body]

If an LV is clustered, it must be deactivated before activating excl.  So,
we move the deactivate so an LV is deactivated regardless.  Then if
clustered, we activate excl (to ensure no-one else has it open/active),
and deactivate.

This should save us crappy messages like:
[root@bp-xen-01 ~]# lvchange --resync vg/lv1
Do you really want to deactivate logical volume lv1 to resync it? [y/n]: y
  Error locking on node bp-xen-01: Volume is busy on another node
  Can't get exclusive access to clustered volume lv1

Index: LVM2-cvs/tools/lvchange.c
===================================================================
--- LVM2-cvs.orig/tools/lvchange.c
+++ LVM2-cvs/tools/lvchange.c
@@ -221,15 +221,18 @@ static int lvchange_resync(struct cmd_co
 		}
 	}
 
+	if (!deactivate_lv(cmd, lv)) {
+		log_error("Unable to deactivate %s for resync", lv->name);
+		return 0;
+	}
+
 	if (vg_is_clustered(lv->vg) && !activate_lv_excl(cmd, lv)) {
 		log_error("Can't get exclusive access to clustered volume %s",
 			  lv->name);
 		return ECMD_FAILED;
-	}
-
-	if (!deactivate_lv(cmd, lv)) {
+	} else if (!deactivate_lv(cmd, lv)) {
 		log_error("Unable to deactivate %s for resync", lv->name);
-		return 0;
+		return ECMD_FAILED;
 	}
 
 	log_lv = first_seg(lv)->log_lv;


[prev in list] [next in list] [prev in thread] [next in thread]