Merge lp:~jk0/nova/lp688545 into lp:~hudson-openstack/nova/trunk

[Josh Kearney]

Fixes LP688545.

Check for a running data store instance upon startup once every 10 seconds, up to 12 times (for a duration of two minutes). Also includes some minor PEP8 cleanup and a prior merge conflict fix.

[Vish Ishaya (vishvananda)]

lgtm. Might be nice to have an exponential falloff. Do we have another bug for recovering from a failed connection?

[Josh Kearney (jk0)]

> lgtm. Might be nice to have an exponential falloff. Do we have another bug
> for recovering from a failed connection?

I *thought* I saw another bug related to that. I'll go back and make a new bug if there isn't.

[Josh Kearney (jk0)]

> lgtm. Might be nice to have an exponential falloff. Do we have another bug
> for recovering from a failed connection?

Turns out I was thinking of the XenAPI session. I've opened bug #697019. I'll address it there tomorrow.

[Todd Willey (xtoddx)]

lgtm

review: Approve

[Vish Ishaya (vishvananda)]

ok looks good.

review: Approve

=== modified file 'nova/compute/manager.py'
--- nova/compute/manager.py	2010-12-30 22:06:48 +0000
+++ nova/compute/manager.py	2011-01-03 22:16:09 +0000
@@ -327,6 +327,7 @@
                 instance_ref["internal_id"])
             return self.driver.get_diagnostics(instance_ref)
 
+    @exception.wrap_exception
     def suspend_instance(self, context, instance_id):
         """suspend the instance with instance_id"""
         context = context.elevated()

=== modified file 'nova/db/sqlalchemy/__init__.py'
--- nova/db/sqlalchemy/__init__.py	2010-08-31 23:48:41 +0000
+++ nova/db/sqlalchemy/__init__.py	2011-01-03 22:16:09 +0000
@@ -19,6 +19,25 @@
 """
 SQLAlchemy database backend
 """
+import logging
+import time
+
+from sqlalchemy.exc import OperationalError
+
+from nova import flags
 from nova.db.sqlalchemy import models
 
-models.register_models()
+
+FLAGS = flags.FLAGS
+
+
+for i in xrange(FLAGS.sql_max_retries):
+    if i > 0:
+        time.sleep(FLAGS.sql_retry_interval)
+
+    try:
+        models.register_models()
+        break
+    except OperationalError:
+        logging.exception(_("Data store is unreachable."
+            " Trying again in %d seconds.") % FLAGS.sql_retry_interval)

=== modified file 'nova/flags.py'
--- nova/flags.py	2011-01-03 19:49:45 +0000
+++ nova/flags.py	2011-01-03 22:16:09 +0000
@@ -266,6 +266,8 @@
 DEFINE_string('sql_idle_timeout',
               '3600',
               'timeout for idle sql database connections')
+DEFINE_integer('sql_max_retries', 12, 'sql connection attempts')
+DEFINE_integer('sql_retry_interval', 10, 'sql connection retry interval')
 
 DEFINE_string('compute_manager', 'nova.compute.manager.ComputeManager',
               'Manager for compute')

=== modified file 'nova/tests/test_xenapi.py'
--- nova/tests/test_xenapi.py	2010-12-31 01:07:20 +0000
+++ nova/tests/test_xenapi.py	2011-01-03 22:16:09 +0000
@@ -250,15 +250,16 @@
 
     def _create_instance(self):
         """Creates and spawns a test instance"""
-        values = {'name': 1, 'id': 1,
-                  'project_id': self.project.id,
-                  'user_id': self.user.id,
-                  'image_id': 1,
-                  'kernel_id': 2,
-                  'ramdisk_id': 3,
-                  'instance_type': 'm1.large',
-                  'mac_address': 'aa:bb:cc:dd:ee:ff'
-                  }
+        values = {
+            'name': 1,
+            'id': 1,
+            'project_id': self.project.id,
+            'user_id': self.user.id,
+            'image_id': 1,
+            'kernel_id': 2,
+            'ramdisk_id': 3,
+            'instance_type': 'm1.large',
+            'mac_address': 'aa:bb:cc:dd:ee:ff'}
         instance = db.instance_create(values)
         self.conn.spawn(instance)
         return instance