commit d4398f715f098b9edbc0be8612b6c079a8e607af
Author: Eric Fried <email address hidden>
Date: Tue Nov 7 09:30:21 2017 -0600
Test alloc candidates with same RC in cn & shared
This change set adds a couple of failing test cases that demonstrate
holes in the design of GET /allocation_candidates when inventory from
the same resource class is present on both the compute node (the "main"
resource provider) and a shared resource provider.
The example being used is where the compute node has some local disk,
and is also associated with a shared storage pool. Both the compute
node RP and the shared storage RP will provide inventory of DISK_GB.
Test case test_common_rc demonstrates bug #1724613: when I ask for
DISK_GB in this setup, the shared storage pool is ignored. I expect to
get two candidates back: one with the storage from the compute node; the
other with the storage from the shared storage pool. But I actually
only get the former candidate back.
Test case test_common_rc_traits_split shows bug #1724633: that placement
can't tell which traits are supposed to apply to which resources. In
the above scenario, if the local storage is SSD and the shared storage
is RAID, and I ask for SSD + RAID, I "expect" to get back no hits. But
I would in fact get back a candidate with the storage from the shared
storage pool, because the cumulative set of traits would satisfy my
requested SSD + RAID.
Note that the two tests are functionally identical (traits are ignored
entirely) until https://review.openstack.org/#/c/479766/ lands. At that
point, depending on how we decide to implement the code that would deal
with this scenario, the test may fail *differently* until bug #1724613
is fixed.
Reviewed: https:/ /review. openstack. org/513149 /git.openstack. org/cgit/ openstack/ nova/commit/ ?id=d4398f715f0 98b9edbc0be8612 b6c079a8e607af
Committed: https:/
Submitter: Zuul
Branch: master
commit d4398f715f098b9 edbc0be8612b6c0 79a8e607af
Author: Eric Fried <email address hidden>
Date: Tue Nov 7 09:30:21 2017 -0600
Test alloc candidates with same RC in cn & shared
This change set adds a couple of failing test cases that demonstrate candidates when inventory from
holes in the design of GET /allocation_
the same resource class is present on both the compute node (the "main"
resource provider) and a shared resource provider.
The example being used is where the compute node has some local disk,
and is also associated with a shared storage pool. Both the compute
node RP and the shared storage RP will provide inventory of DISK_GB.
Test case test_common_rc demonstrates bug #1724613: when I ask for
DISK_GB in this setup, the shared storage pool is ignored. I expect to
get two candidates back: one with the storage from the compute node; the
other with the storage from the shared storage pool. But I actually
only get the former candidate back.
Test case test_common_ rc_traits_ split shows bug #1724633: that placement
can't tell which traits are supposed to apply to which resources. In
the above scenario, if the local storage is SSD and the shared storage
is RAID, and I ask for SSD + RAID, I "expect" to get back no hits. But
I would in fact get back a candidate with the storage from the shared
storage pool, because the cumulative set of traits would satisfy my
requested SSD + RAID.
Note that the two tests are functionally identical (traits are ignored /review. openstack. org/#/c/ 479766/ lands. At that
entirely) until https:/
point, depending on how we decide to implement the code that would deal
with this scenario, the test may fail *differently* until bug #1724613
is fixed.
Related-Bug: #1724613
Related-Bug: #1724633
Change-Id: I42edf102379cf3 29aa2252ab779a9 f945f5fc155