commit c3eb28078ab12f908c2536bf78f68c0ee0d4ac0a
Author: Amit Saha <asaha@redhat.com>
Date:   Thu Aug 14 09:12:29 2014 +1000

    Automatic commit of release 0.17.3

commit 860fb557cdcb4dd6d26128a412b56e1bf1c3d162
Author: Dan Callaghan <dcallagh@redhat.com>
Date:   Mon Aug 11 20:08:16 2014 +1000

    release notes for 0.17.3
    
    Change-Id: I2495e7b5ef841b7b95b792f79eafeea281d8d4ee

commit 4ced84a7363b1a29f44231554092f5df8980024d
Author: Dan Callaghan <dcallagh@redhat.com>
Date:   Mon Aug 11 14:40:06 2014 +1000

    set netboot_method= for GRUB2
    
    Change-Id: Ia748f219af045aa6709ceef16cdb32a21584c917

commit 491af1c7eeb358c15e77977de544a434484626a8
Author: Dan Callaghan <dcallagh@redhat.com>
Date:   Fri Aug 8 09:58:53 2014 +1000

    fix mistake in db downgrade notes for 0.17
    
    We need to add back the lab_controller.user_id index in the same ALTER
    TABLE statement, because the column is involved in an InnoDB foreign key
    and InnoDB requires an index on it.
    
        ERROR 1553 (HY000): Cannot drop index 'uc_user_id': needed in a foreign key constraint
    
    Change-Id: Idff5fce194fca70c58a6dbf22cf048635895e8e2

commit c8c365d6fabb9e85d29bd8e124cc8565cbaeaab2
Author: Matt Jia <mjia@redhat.com>
Date:   Wed Aug 6 15:34:29 2014 +1000

    Fix for regression that recipe-task can not be cancelled by CLI
    
    bug: 1124756
    Change-Id: Ic1f46e75137e2c81c76f26553f2e42174ef39024

commit 636ff78e20c5cc07d8e78ea32de52b42978ba685
Author: Amit Saha <asaha@redhat.com>
Date:   Fri Aug 1 00:24:57 2014 +1000

    Support for PPC64 grub2 netboot
    
    Bug: 1113816
    Change-Id: Icb08cbfe44ad85e8c4a34a645a77e3a1bcf2361a

commit c48d0e4edbf8e6cae125c4c23ee770c301325c89
Author: Dan Callaghan <dcallagh@redhat.com>
Date:   Mon Aug 4 17:33:33 2014 +1000

    prevent duplicate users due to LDAP whitespace normalization
    
    Bug: 1121748
    Change-Id: I3b6563295699414ac867c1c14f0b3b9bd4addf83

commit 91464919cd80e74061a25f8f72eca34d7e5f875c
Author: Dan Callaghan <dcallagh@redhat.com>
Date:   Wed Jul 23 09:53:20 2014 +1000

    abort the entire worker process if session.close() fails
    
    If we fail to close the session (for example because of an out-of-memory
    condition) we want the entire worker process to abort, because any
    subsequent requests will fail due to the transaction left open. Under
    mod_wsgi daemon mode the worker will just be replaced with a fresh one.
    
    Also removed superfluous session.rollback() call from the close_session
    handler. The commit_or_rollback_session handler would already have tried
    to call session.rollback(), so if that failed for some reason then
    trying again will only make things worse. session.close() will
    ultimately send a ROLLBACK to the database as part of releasing the
    connection anyway, which is really what we care about.
    
    Bug: 1120439
    Change-Id: Ibd6cd6deb924a9e91f586ff841c4dd24859131ec

commit 7abb79ad76be54c8443314eae87d661ef967a148
Author: Amit Saha <asaha@redhat.com>
Date:   Tue Jul 29 11:30:52 2014 +1000

    ppc64le: Add to the list of architectures
    
    Bug: 1120487