Reviewed:  https://review.openstack.org/538415
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=0e9cd6c4d66ca4afb95bb60edb412af9e96c546e
Submitter: Zuul
Branch:    master

commit 0e9cd6c4d66ca4afb95bb60edb412af9e96c546e
Author: Brooks Kaminski <email address hidden>
Date:   Sat Jan 27 01:35:07 2018 -0600

    XenAPI: XCP2.1+ Swallow VDI_NOT_IN_MAP Exception
    
    Changes within XenAPI have enforced a more strict policy when checking
    assert_can_migrate.  In particular when checking the source_vdi:dest_sr
    mapping it insists that the SR actually exist.  This is not a problem for
    local disks, however this assertation is called extremely early in the
    live migration process (check_can_migrate_source) which is called from
    conductor, which makes a problem for attached volumes.
    
    This early in the process the host has just barely been chosen and no SR
    information has been configured yet for these volumes or their initiators.
    Additionally we cannot prepare this SR any earlier as BDM information is
    not set up until the pre_live_migration method. With the options to either
    skip this assertion completely or swallow the exception, I have chosen to
    swallow the exception.  My reasons for this are two-fold:
    
    1. --block-migration can be called without regard for whether an iSCSI
    volume is attached, and we still want to ensure that VIF, CPU and other
    factors are checked, and not just skip all checks entirely.
    2. Currently the Assert only exists within the --block-migration code
    base but this needs to change. A future commit will remove this logic
    to ensure that the commit runs without this flag. Once that is done we
    want to be able to continue to use this Exception swallow logic rather
    than continuing to skip the assert for all XCP2.1.0+ even without volumes.
    
    This decision should help us handle less work in a future commit and does not
    seem to align with the goals of that commit, where it does align properly here.
    This commit still changes very little of the current codebase and puts us in
    a good position to refactor the way this is handled at a later date, while
    adding a TODO note to correct VM.assert_can_migrate only running during a
    block migration.
    
    Additionally there seems to be some confusion that the mapping data that is
    generated during this initial trip through _call_live_migrate_command is needed
    to continue along the code, however this data appears to be purely used to send
    the mapping information through the assertation call, and is then discarded.
    The only data returned from these methods is the original dest_data which
    is carried into the live_migration method. The _call_live_migration method is
    called again during the live_migration() method, and during this time it does
    need that mapping to send along to XenAPI for the actual migration, but not
    yet. Because this codebase is so confusing, I am providing a little bit of
    context on the movement of these variables with some psuedocode:
    
    ---CONDUCTOR.TASKS.LIVE_MIGRATE---
    LiveMigrationTask.Execute()
    	self._find_destination() <-----
    	Unrelated Work
    	compute.live_migration(self, host, instance, destination,
    			       block_migrate, migration, migrate_data)
    
    LiveMigrationTask._find_destination()
    	Scheduler Things.  Gets a Dest ref.
    	_check_compatible_with_source_hyp
    	_call_livem_checks_on_host(host) <-----
    		_check_can_live_migrate_destination()
    	returns Host Node Name and Host Compute.  That's all.
    
    ---COMPUTE.MANAGER---
    _do / _check_live_migration_destination
    	dest_check_data = xenops.can_live_migrate_destination
    
    	(Checks for the Assert)
    	try:
    		migrate_data = check_can_live_migrate_source(dest_check_data)
    
    	return migrate_data
    
    ---VMOPS--
    check_can_migrate_source(self, ctxt, instance_ref, dest_check_data)
    	if block_migration:
    		_call_live_migration_command(assert_can_migrate)
    			_generate_vdi_map()
    				Does NOT return
    			ALSO Does NOT return
    return dest_check_data
    
    The changes made to address this issue are a fairly simple oslo_utils
    version check. To pull this data I created two new host related methods
    within VMops as well as a new import of oslo_config.versionutils.
    I believe these methods ultimately belong in the xenapi.host class, but
    for two very small methods I believed it better to avoid such a large import
    to get minimal data.
    
    Finally, an adjustment to the Fake XenAPI driver had to be made as it currently
    does not include the host details beyond hostname environment in the
    create_host check.  The change amends the stub dictionary to include this data.
    
    Change-Id: Ie5295564a68f877fc061376c00f3fa706370a251
    Closes-Bug: #1704071