placement: use simple code paths when possible

Somewhere in the past release, we started using extremely complex code
paths involving sharing providers, anchor providers, and nested resource
provider calculations when we absolutely don't need to do so.

There was a _has_provider_trees() function in the
nova/api/openstack/placement/objects/resource_provider.py file that used
to be used for top-level switching between a faster, simpler approach to
finding allocation candidates for a simple search of resources and
traits when no sharing providers and no nesting was used. That was
removed at some point and all code paths -- even for simple "get me
these amounts of these resources" when no trees or sharing providers are
present (which is the vast majority of OpenStack deployments) -- were
going through the complex tree-search-and-match queries and algorithms.

This patch changes that so that when there's a request for some
resources and there's no trees or sharing providers, we do the simple
code path. Hopefully this gets our performance for the simple, common
cases back to where we were pre-Rocky.

This change is a prerequisite for the following change which adds
debugging output to help diagnose which resource classes are running
out of inventory when GET /allocation_candidates returns 0 results.
That code is not possible without the changes here as they only
work if we can identify when a "simpler approach" is possible and
call that simpler code.

Related-Bug: #1786055
Partial-Bug: #1786519
Change-Id: I1fdbcdb7a1dd51e738924c8a30238237d7ac74e1
This commit is contained in:
Jay Pipes 2018-08-09 10:46:20 -04:00
parent 327448e7cb
commit 9ea340eb0d
1 changed files with 25 additions and 17 deletions

View File

@ -3976,7 +3976,7 @@ class AllocationCandidates(base.VersionedObject):
)
@staticmethod
def _get_by_one_request(context, request):
def _get_by_one_request(context, request, sharing_providers, has_trees):
"""Get allocation candidates for one RequestGroup.
Must be called from within an placement_context_manager.reader
@ -3984,6 +3984,12 @@ class AllocationCandidates(base.VersionedObject):
:param context: Nova RequestContext.
:param request: One nova.api.openstack.placement.util.RequestGroup
:param sharing_providers: dict, keyed by resource class internal ID, of
the set of provider IDs containing shared
inventory of that resource class
:param has_trees: bool indicating there is some level of nesting in the
environment (if there isn't, we take faster, simpler
code paths)
:return: A tuple of (allocation_requests, provider_summaries)
satisfying `request`.
"""
@ -4006,26 +4012,14 @@ class AllocationCandidates(base.VersionedObject):
missing = traits - set(trait_map)
raise exception.TraitNotFound(names=', '.join(missing))
# Microversions prior to 1.21 will not have 'member_of' in the groups.
# This allows earlier microversions to continue to work.
member_of = getattr(request, "member_of", None)
member_of = request.member_of
if not request.use_same_provider:
any_sharing = any(sharing_providers.values())
if not request.use_same_provider and (has_trees or any_sharing):
# TODO(jaypipes): The check/callout to handle trees goes here.
# Build a dict, keyed by resource class internal ID, of lists of
# internal IDs of resource providers that share some inventory for
# each resource class requested.
# TODO(jaypipes): Consider caching this for some amount of time
# since sharing providers generally don't change often and here we
# aren't concerned with how *much* inventory/capacity the sharing
# provider has, only that it is sharing *some* inventory of a
# particular resource class.
sharing_providers = {
rc_id: _get_providers_with_shared_capacity(context, rc_id,
amount, member_of)
for rc_id, amount in resources.items()
}
# If there aren't any providers that have any of the
# required traits, just exit early...
if required_trait_map:
@ -4064,9 +4058,23 @@ class AllocationCandidates(base.VersionedObject):
@db_api.placement_context_manager.writer
def _get_by_requests(cls, context, requests, limit=None,
group_policy=None):
# TODO(jaypipes): Make a RequestGroupContext object and put these
# pieces of information in there, passing the context to the various
# internal functions handling that part of the request.
sharing = {}
for request in requests.values():
member_of = request.member_of
for rc_name, amount in request.resources.items():
rc_id = _RC_CACHE.id_from_string(rc_name)
if rc_id not in sharing:
sharing[rc_id] = _get_providers_with_shared_capacity(
context, rc_id, amount, member_of)
has_trees = _has_provider_trees(context)
candidates = {}
for suffix, request in requests.items():
alloc_reqs, summaries = cls._get_by_one_request(context, request)
alloc_reqs, summaries = cls._get_by_one_request(
context, request, sharing, has_trees)
LOG.debug("%s (suffix '%s') returned %d matches",
str(request), str(suffix), len(alloc_reqs))
if not alloc_reqs: