This reverts commit 4b142e53e43132e996d35c335da30959f9f361be.
I was having a point left unresolved in https://review.openstack.org/#/c/189279/8/doc/source/filter_scheduler.rst about the tech debt that adding a new filter in-tree would cause for something very closely related to another filter.
I totally get the need for adding more logic and reverting how we compare flavors vs. aggregates. I just feel that before committing ourselves to that, we need to correctly estimate the possibilities to modify AggregateInstanceExtraSpecsFilter to fit the above needs.
Change-Id: Ic9f70dae037d32980a5a252bdd08eff02ba27120
By default this is turned on and has equal weight to the
ram weigher.
DocImpact: Added the disk_weight_multiplier config option
Closes-Bug: 1513654
Change-Id: I29ce73122ad1860081b64b75646a297dfbb8d292
The flavor_extra_spec metadata pair will be consumed by the
AggregateTypeExtraSpecsAffinityFilter to allow operators to define a
set of extra specs key value pairs that are required to schedule to
the aggregate, e.g.:
standard memory backing aggregate:
flavor_extra_spec: "hw:mem_page_size=small,hw:mem_page_size=any"
high bandwidth memory backing aggregate:
flavor_extra_spec: "hw:mem_page_size=2M,hw:mem_page_size=1G," \
"hw:mem_page_size=large"
DocImpact
Implements: blueprint aggregate-extra-specs-filter
Change-Id: Id3a9918cf9f83b2a9b1dfbcd91803b5b1b2bcc78
The versioned notificaton infrastructure is in place and the team
agreed on the midcycle to allow new notification types only based
on the versioned notification infrastructure. So the code-review
document is updated to reflect this decision
Change-Id: I7f3860ca59658dc21e67f26bb45341cb46da2235
Partially-Implements: bp versioned-notification-api
This commit adds a new sphinx extension that inspects the nova code
and adds information about the existing versioned notifications to
the nofitications devref. This way the devref is automatically kept
up to date with relevant information.
Partially-Implements: bp versioned-notification-api
Change-Id: If65d5d81e26cb2b4a9f57a8c7d37d3de310ebe00
Post Ic1f106684270a8f2197e0fb2a67d6fc00535ee65 we now require
msgfmt (a binary in the gettext package) to be installed for pep8
to pass. This hasn't failed in the gate because infra already
installs that package. We should make sure new devs know to do
that thing as well.
Change-Id: Iac505ddd54466e8a6272fee5ff61aac180c90ac4
This patch update feature support matrix documentation for uefi boot.
Change-Id: I8ec6a2ad24e100cce48d70160350e5bf853992ef
Implements: blueprint boot-from-uefi
Add the native distro packages for python 3 development to the dev
quickstart guide (needed for running the tox py34 target).
Change-Id: Iae651ac2f7d4b050cba46ee27e5a6d4232f3fddc
Since I8ad9313f824ef656c841a9b60ee575f02ddea321, the api-wg has
introduced a new guideline. That was based on Nova's development
experience, and this patch adds it as a reviewing point for making
consistent APIs.
Change-Id: I53c4055d973b6f76fca16ae170d0df9e1829fca5
In Id7936be290b6febd18deb4c2db8ea4d678d4d9b1, we removed
entries from api-paste.ini for EC2 API service. In this
review we remove all the unnecessary code, docs and tests
associated with objectstore and ec2 service. Note that this
does not cleanup the Instance object or change any of the
versioned objects. We just drop any code associated with
testing the REST endpoint(s) that are no longer needed.
Also added shims such that the api-paste.ini from liberty
will still work (grenade job) and added logs and response
messages for prompting administrators to cleanup their
old api-paste.ini and switching to the stand alone EC2 API
project for their needs.
Change-Id: I8bf7cbaa7015bb61656ab90ccc8f944aaeebb095
ServerGroupSoftAffinityWeigher implements the soft-affinity
policy for server groups as it orders the hosts by the number
of instances running on them from the same group in decreasing
order.
ServerGroupSoftAntiAffinityWeigher implements the
soft-anti-affinity policy for server groups as it orders the
hosts by the number of instances running on them from the same
group in increasing order.
Both weigher assumes the the RequestSpec object refers to a valid
InstanceGroup object which has a populated members field. This
will be provided by the subsequent patch.
New configuration options added:
* soft_affinity_weight_multiplier
* soft_anti_affinity_multiplier
Add docs about the new weigher classes and change the function
name 'weigh_object' to '_weigh_object' in the Weights section
to match the code.
DocImpact
Co-Authored-By: Balazs Gibizer <balazs.gibizer@ericsson.com>
Implements: blueprint soft-affinity-for-server-group
Change-Id: I3f156d5e5df4d9642bb4b0ffac30a6288459ce61
os.popen() is deprecated since version 2.6. Resolved with use of
subprocess module.
Change-Id: Iea9c0501f46c57d809217912fb0f40eb2bc4f111
Closes-Bug: #1529836
The trivial bug list has already got a +1 and often the fixes there are
too trivial to learn very much. Replaced with some hopefully more
useful advice.
Change-Id: Ibf5abc4dca170770d302ae9276312faad5eabdb6
Make it easier to add additional support-matrix entries by making the
filename of the ini file a required argument of the directive.
As a side effect, its clearer in the rst file where the data is stored.
Change-Id: I05bd30b64152368b53c030a32754d0ed4a35bf88
Based on https://review.openstack.org/#/c/214691/
Some concern raised about remove a header may need
microversion bump, update devref to indicate that
it sould be ok to update without microversion usage.
Co-Authored-By: Alexis Lee <lxsli@hpe.com>
Change-Id: I0b9e6e384b54ae8bfc7896f0135cd443ccfe2b1e
Basic conversion done with pandoc then I had to fix some links and the
title underlines.
The Mentoring page has been merged into the (mostly empty) "How to get
(more) involved with Nova" page.
Change-Id: Ib4a11b21c0fd6aaf86e643897193f59cefb626c4
Since the 2015/11/12 nova meeting we've said that cells v1 is in feature
freeze and we won't focus on hard-to-fix latent bugs. We should include that
information in the devref so it's treated as a policy.
Change-Id: I31186463fd832c27accc05feb82713fe71f603a8
The config option scheduler_default_filters affects scheduling
performance, This commit explains what we should consider to
get better scheduling performance.
Change-Id: I7cd17cbcf037a9fb6c5389b1005c0c2501fe22f2
With blueprint centralize-config-options we came to a conclusion
that it makes sense to have all config options at one central place.
This patch enhances the "config options" section with a "location"
subsection to clarify this.
Change-Id: I48b6f0c3311ad75f9faef78a7d163ae5bf640c6c
Found another odd reference to XML in the api while grepping
around. This should just be removed, it's only confusing to people.
Change-Id: I7c4d9dc69fb386d937197ef466a6e96495dd8045
Replace 'weight_object' with '_weight_object' in doc
and add more details about how to implement subclass
of weights.BaseHostWeigher.
Change-Id: Ia15c04fdd07dbb19c08051b721eb822a244f8986
There are two ways to give other greenthread chance to run:
greenthread.sleep(0) or time.sleep(0). Add the second way in
threading.rst and recommend the second way for contributors.
Change-Id: I33dc73a0edfa5fde0e5fa8051a594874ab166970
Jenkins