Instance migration: add delays around migration

It seems that xen can have issues that are triggered by too fast
migrations. For now, until we have more experience with it, we should
add some delays before/after the migration call so that things have
enough time to settle down (e.g. hot plug scripts, etc.).

Note that I don't have solid data that this will fix it, or that indeed
xen is the culprit, but for now this seems the simplest way to try to
mitigate it.

Reviewed-by: ultrotter

diff --git a/lib/cmdlib.py b/lib/cmdlib.py
index edea930..9d5e23d 100644 (file)
--- a/lib/cmdlib.py
+++ b/lib/cmdlib.py
@@ -3160,12 +3160,12 @@ class LUMigrateInstance(LogicalUnit):
     self._WaitUntilSync()
 
     self.feedback_fn("* migrating instance to %s" % target_node)
+    time.sleep(2)
     result = rpc.call_instance_migrate(source_node, instance,
                                        self.nodes_ip[target_node],
                                        self.op.live)
     if not result or not result[0]:
       logger.Error("Instance migration failed, trying to revert disk status")
-
       try:
         self._EnsureSecondary(target_node)
         self._GoStandalone()
@@ -3177,6 +3177,7 @@ class LUMigrateInstance(LogicalUnit):
 
       raise errors.OpExecError("Could not migrate instance %s: %s" %
                                (instance.name, result[1]))
+    time.sleep(2)
 
     instance.primary_node = target_node
     # distribute new instance config to the other nodes