Posted
over 5 years
ago
by
Boris Renski
With edge, we have closed vendor solutions and open reference architectures, but nothing for VNF and software vendors to build on top of. Changing that is the first step towards virtualized edge.
|
Posted
over 5 years
ago
by
Gary Kevorkian
With just over two weeks to go before the OpenStack Summit EU, the Cisco team is starting to get very excited about being back at this awesome event. For...
|
Posted
over 5 years
ago
by
Superuser
Highlights you’ll want to add to your schedule about HPC, GPU and AI.
The post Inside HPC, GPU, AI : Must-see sessions at the Berlin Summit appeared first on Superuser.
|
Posted
over 5 years
ago
by
Nicole Martinelli
Join the people building and operating open infrastructure at the OpenStack Summit Berlin in November. The Summit schedule features over 200 sessions organized by use cases including: artificial intelligence and machine learning, high performance
... [More]
computing, edge computing, network functions virtualization, container infrastructure and public, private and multi-cloud strategies. Here we’re highlighting some of the sessions […]
The post Inside HPC, GPU, AI : Must-see sessions at the Berlin Summit appeared first on Superuser.
[Less]
|
Posted
over 5 years
ago
by
Chris Dent
A placement update for you.
Most Important
Same as last week: The major factors that need attention are
managing database migrations and associated tooling and getting the
ball rolling on properly producing documentation. More on both of
these things
... [More]
in the extraction section below.
Matt has sent out an
email
seeking volunteers from OpenStack Ansible or TripleO to get
placement upgrade tooling in one of those projects.
Bugs
I guess it is because of various people doing upgrades, and some of
the downstream projects starting to take more advantage of
placement, but there's been a raft of interesting bugs recently.
Many related to some of the more esoteric aspects of the
ProviderTree handling in the resource tracker, the SQL in the
placement service, or management of global state in WSGI servers.
Initially this is a bit frustrating, but it's also a good thing:
Finding and fixing bugs is the beating heart of an open source
project. So thanks to everyone finding and fixing them.
Placement related bugs not yet in progress: 15.
-1.
In progress placement bugs 11.
Specs
The spec review sprint happened and managed to get some specs
merged, so this list should have shrunk some.
https://review.openstack.org/#/c/544683/
Account for host agg allocation ratio in placement
(Still in rocky/)
https://review.openstack.org/#/c/595236/
Add subtree filter for GET /resource_providers
https://review.openstack.org/#/c/597601/
Resource provider - request group mapping in allocation candidate
https://review.openstack.org/#/c/549067/
VMware: place instances on resource pool
(still in rocky/)
https://review.openstack.org/#/c/555081/
Standardize CPU resource tracking
https://review.openstack.org/#/c/599957/
Allow overcommit of dedicated CPU
(Has an alternative which changes allocations to a float)
https://review.openstack.org/#/c/591037/
Modelling passthrough devices for report to placement
https://review.openstack.org/#/c/603585/
Spec: allocation candidates in tree
https://review.openstack.org/#/c/603805/
[WIP] generic device discovery policy
https://review.openstack.org/#/c/603955/
Nova Cyborg interaction specification.
https://review.openstack.org/#/c/601596/
supporting virtual NVDIMM devices
https://review.openstack.org/#/c/603352/
Spec: Support filtering by forbidden aggregate
https://review.openstack.org/#/c/552924/
Proposes NUMA topology with RPs
https://review.openstack.org/#/c/569011/
Count quota based on resource class
https://review.openstack.org/#/c/607989/
WIP: High Precision Event Timer (HPET) on x86 guests
https://review.openstack.org/#/c/571111/
Add support for emulated virtual TPM
https://review.openstack.org/#/c/141219/
Adds spec for instance live resize
https://review.openstack.org/#/c/612497/
Provider config YAML file
Main Themes
Making Nested Useful
Work on getting nova's use of nested resource providers happy and
fixing bugs discovered in placement in the process. This is creeping
ahead. We recently confirmed that end-to-end success with nested
providers is priority one for resource provider related work.
https://review.openstack.org/#/q/topic:bp/use-nested-allocation-candidates
There's a topic for reshaper that still has some open patches:
https://review.openstack.org/#/q/topic:bp/reshape-provider-tree+status:open
Extraction
There continue to be three main tasks in regard to placement
extraction:
upgrade and integration testing
database schema migration and management
documentation publishing
There's been some good progress here. The grenade job
works and is ready to
merge independent of other things. The related devstack
change is still waiting
on the database management that's part of (2). As mentioned above,
volunteers from OSA or TripleO are being recruited.
That db management is making some good headway with a working
alembic setup but the
tooling to use it needs to be formalized. The command line
hack has been updated to
use the alembic setup.
We have work in progress to tune up the documentation but we are not
yet publishing documentation (3). The plan here is to incrementally
improve things as we have attention and discover things. One goal
with this is to keep the process moving and use followups to avoid
nitpicking each other too much.
Other
Various placement changes out in the world.
https://review.openstack.org/#/c/601866/
Generate sample policy in placement directory
(This is a bit stuck on not being sure what the right thing to do
is.)
https://review.openstack.org/#/q/topic:bp/initial-allocation-ratios
Improve handling of default allocation ratios
https://review.openstack.org/#/q/topic:minimum-bandwidth-allocation-placement-api
Neutron minimum bandwidth implementation
https://review.openstack.org/#/c/602160/
Add OWNERSHIP $SERVICE traits
https://review.openstack.org/#/c/604182/
Puppet: Initial cookiecutter and import from nova::placement
https://review.openstack.org/#/c/586960/
zun: Use placement for unified resource management
https://review.openstack.org/#/q/topic:bug/1799727
Update allocation ratio when config changes
https://review.openstack.org/#/q/topic:bug/1799892
Deal with root_id None in resource provider
https://review.openstack.org/#/q/topic:bug/1795992
Use long rpc timeout in select_destinations
https://review.openstack.org/#/c/529343/
Cleanups for scheduler code
https://review.openstack.org/#/q/topic:bp/bandwidth-resource-provider
Bandwith Resource Providers!
https://review.openstack.org/#/q/topic:bug/1799246
Harden placement init under wsgi
https://review.openstack.org/#/q/topic:cd/gabbi-tempest-job
Using gabbi-tempest for integration tests.
https://review.openstack.org/#/c/613118/
Make tox -ereleasenotes work
https://review.openstack.org/#/c/613343/
placement: Add a doc describing a quick live environment
End
It's tired around here.
[Less]
|
Posted
over 5 years
ago
by
Trinh Nguyen
|
Posted
over 5 years
ago
by
Superuser
Answers to your most common questions about the upcoming Forum at the Summit Berlin.
The post What to expect from the upcoming OpenStack Forum appeared first on Superuser.
|
Posted
over 5 years
ago
by
Chris Dent
One day a while back, I started blabbing out loud about some quick
ways to experiment with a live placement service and was (very
appropriately) reminded that I make a ton of incorrect assumptions
about the familiarity people have with things I do
... [More]
most days.
Statements like "just spin up the wsgi script against uwsgi and a
stubbed out placement.conf and see what happens" are a good example
of me being bad.
So, to help, here are some instructions on how to spin up the wsgi
script against uwsgi and a stubbed out placement.conf, in case you
want to see what happens. The idea here is that you want to
experiment with the current placement code, using a live database,
but you're not concerned with other services, don't want to deal
with devstack, but need a level of interaction with the code and
process that something like
placedock can't provide.
As ever, even all of the above has lots of assumptions about
experience and context. This post assumes you are someone who either
is an OpenStack (and probably placement) developer, or would like to
be one.
To make this go you need a unix-like OS, with a python3 dev
environment, and git and mysql (or postgresql) installed. We'll be
doing this work from within a virtualenv, built from the tox.ini
in the placement code.
At the time of writing, some required code is not yet merged into
placement, so we'll be using a patch that is currently under review.
I'll update this document when that code merges so we can skip a
step.
Get The Code and Deps
The placement code lives at
https://git.openstack.org/cgit/openstack/placement. We want to
clone that:
git clone git://git.openstack.org/openstack/placement
cd placement
Then we want to get the extra code mentioned above:
git pull https://git.openstack.org/openstack/placement refs/changes/61/600161/13
Setup The Database
That patch adds a commands to create tables in a database. We need
to 1) create the database, 2) create a virtualenv to have the
command, 3) use it to create the tables.
The database can have whatever name you like. Whatever you choose,
use it throughout this process. I choose placement. You may need a
user and password to talk to your database, setting that up is out
of scope for this document. I'm using a machine that has had
devstack on it before so mysql is configured in what might be a
familiar way:
mysql -uroot -psecret -e "DROP DATABASE IF EXISTS placement;"
mysql -uroot -psecret -e "CREATE DATABASE placement CHARACTER SET utf8;"
You may also need to set permissions:
mysql -uroot -psecret \
-e "GRANT ALL PRIVILEGES ON placement.* TO 'root'@'%' identified by 'secret';"
Get the table create command by updating the virtualenv:
tox -epy36 --notest
Create a bare minimum placement.conf in the /etc/placement
directory (which you may need to create):
[placement_database]
connection = mysql+pymysql://root:[email protected]/placement?charset=utf8
(Note that when this command matures you will be able to name the
location of the configuration file on the command line.)
Run the command to create the tables:
.tox/py36/bin/placement-manage db table_create
You can confirm the tables are there with mysqlshow placement
Run The Service
Now we want to run the service. We need to update placement.conf
so it will produce debugging output and use the noauth strategy
for authentication (so we don't also have to run Keystone). Make
placement.conf look like this (adjusting for your database
settings):
[DEFAULT]
debug = True
[placement_database]
connection = mysql+pymysql://root:[email protected]/placement?charset=utf8
[api]
auth_strategy = noauth2
We need to install the uwsgi package into the virtualenv:
.tox/py36/bin/pip install uwsgi
And then use uwsgi to run the service. Start it with:
.tox/py36/bin/uwsgi --http :8000 --wsgi-file .tox/py36/bin/placement-api
If that worked you'll see lots of debug output and spawned uWSGI
worker 1. Test that things are working from another terminal with
curl:
curl -v http://localhost:8000/
Get a list of resource providers with (the x-auth-token header
is required, openstack-api-version is optional but makes sure
we are getting the latest functionality):
curl -H 'x-auth-token: admin' \
-H 'openstack-api-version: placement latest' \
http://localhost:8000/resource_providers
The result ought to look something like this:
{"resource_providers": []}
If it doesn't then something went wrong with the above and there
should be more information in the terminal where uwsgi is running.
From here you can experiment with creating resource providers and
related placement features. If you change the placement code,
ctrl-c to kill the uwsgi process and start it up again. For
testing, you might might enjoy
placecat.
Here's a script to do the install for you:
[Less]
|
Posted
over 5 years
ago
by
Nicole Martinelli
Insight from a recent market research study conducted by ClearPath Strategies and the OpenStack Foundation.
The post Taking a closer look at open source infrastructure appeared first on Superuser.
|
Posted
over 5 years
ago
by
Allison Price
OVH hits the right notes at their annual Summit in Paris.
The post Rocking the innovation revolution with open source appeared first on Superuser.
|