Index: lnt/server/db/migrations/upgrade_12_to_13.py
===================================================================
--- /dev/null
+++ lnt/server/db/migrations/upgrade_12_to_13.py
@@ -0,0 +1,18 @@
+# Version 12 drops the info_key column as it is unused now.
+import sqlalchemy
+
+def upgrade(engine):
+    # Create a session.
+    session = sqlalchemy.orm.sessionmaker(engine)()
+
+    # We do not convert names anymore when importing/exporting.
+    # So the info_keys only live in lnt.testing.upgrade_1_to_2() when upgrading
+    # older reports.
+    if engine.dialect.name != 'sqlite':    # sqlite cannot DROP COLUMNs :-(
+        session.connection().execute("""
+ALTER TABLE TestSuiteMachineFields DROP COLUMN info_key;
+ALTER TABLE TestSuiteOrderFields DROP COLUMN info_key;
+ALTER TABLE TestSuiteRunFields DROP COLUMN info_key;
+ALTER TABLE TestSuiteSampleFields DROP COLUMN info_key;
+                                     """)
+    session.commit()
Index: lnt/server/db/testsuite.py
===================================================================
--- lnt/server/db/testsuite.py
+++ lnt/server/db/testsuite.py
@@ -98,21 +98,15 @@
                            index=True)
     name = Column("Name", String(256))
 
-    # The info key describes the key to expect this field to be present as in
-    # the reported machine information. Missing keys result in NULL values in
-    # the database.
-    info_key = Column("InfoKey", String(256))
-
-    def __init__(self, name, info_key):
+    def __init__(self, name):
         self.name = name
-        self.info_key = info_key
 
         # Column instance for fields which have been bound (non-DB
         # parameter). This is provided for convenience in querying.
         self.column = None
 
     def __repr__(self):
-        return '%s%r' % (self.__class__.__name__, (self.name, self.info_key))
+        return '%s%r' % (self.__class__.__name__, self.name)
 
 class OrderField(FieldMixin, Base):
     __tablename__ = 'TestSuiteOrderFields'
@@ -122,20 +116,14 @@
                            index=True)
     name = Column("Name", String(256))
 
-    # The info key describes the key to expect this field to be present as in
-    # the reported machine information. Missing keys result in NULL values in
-    # the database.
-    info_key = Column("InfoKey", String(256))
-
     # The ordinal index this field should be used at for creating a
     # lexicographic ordering amongst runs.
     ordinal = Column("Ordinal", Integer)
 
-    def __init__(self, name, info_key, ordinal):
+    def __init__(self, name, ordinal):
         assert isinstance(ordinal, int) and ordinal >= 0
 
         self.name = name
-        self.info_key = info_key
         self.ordinal = ordinal
 
         # Column instance for fields which have been bound (non-DB
@@ -143,8 +131,7 @@
         self.column = None
 
     def __repr__(self):
-        return '%s%r' % (self.__class__.__name__, (self.name, self.info_key,
-                                                   self.ordinal))
+        return '%s%r' % (self.__class__.__name__, (self.name, self.ordinal))
 
 class RunField(FieldMixin, Base):
     __tablename__ = 'TestSuiteRunFields'
@@ -154,21 +141,15 @@
                            index=True)
     name = Column("Name", String(256))
 
-    # The info key describes the key to expect this field to be present as in
-    # the reported machine information. Missing keys result in NULL values in
-    # the database.
-    info_key = Column("InfoKey", String(256))
-
-    def __init__(self, name, info_key):
+    def __init__(self, name):
         self.name = name
-        self.info_key = info_key
 
         # Column instance for fields which have been bound (non-DB
         # parameter). This is provided for convenience in querying.
         self.column = None
 
     def __repr__(self):
-        return '%s%r' % (self.__class__.__name__, (self.name, self.info_key))
+        return '%s%r' % (self.__class__.__name__, self.name)
 
 class SampleField(FieldMixin, Base):
     __tablename__ = 'TestSuiteSampleFields'
@@ -182,11 +163,6 @@
     type_id = Column("Type", Integer, ForeignKey('SampleType.ID'))
     type = relation(SampleType)
 
-    # The info key describes the key to expect this field to be present as in
-    # the reported machine information. Missing keys result in NULL values in
-    # the database.
-    info_key = Column("InfoKey", String(256))
-
     # The status field is used to create a relation to the sample field that
     # reports the status (pass/fail/etc.) code related to this value. This
     # association is used by UI code to present the two status fields together.
@@ -198,11 +174,10 @@
     # This assumption can be inverted by setting this column to nonzero.
     bigger_is_better = Column("bigger_is_better", Integer)
     
-    def __init__(self, name, type, info_key, status_field = None,
+    def __init__(self, name, type, status_field = None,
                  bigger_is_better = 0):
         self.name = name
         self.type = type
-        self.info_key = info_key
         self.status_field = status_field
         self.bigger_is_better = bigger_is_better
 
@@ -214,5 +189,4 @@
         self.column = None
 
     def __repr__(self):
-        return '%s%r' % (self.__class__.__name__, (self.name, self.type,
-                                                   self.info_key))
+        return '%s%r' % (self.__class__.__name__, (self.name, self.type))