Omnitruck makes it easy to download and install omnibus artifacts. It provides an API to query the available versions of artifacts, get detailed information about the available versions and download them.
Powered by Sinatra.
Renders the install script for Chef packages.
mixlib-install
is used under to covers to generate the scripts. The base url for the scripts will be equivalent to the base url of the running omnitruck instance.
Returns package information based on the provided parameters. Returned information can be json
or txt
and it looks like:
json:
{
"url": "...",
"sha1": "...",
"sha256": "...",
"version": "..."
}
text:
url\t...\nsha1\t...\nsha256\t...\nversion\t...\n
Supported parameters are:
Description | |
---|---|
v |
Version string. It can be full (12.9.2) or partial (12). If omitted, the latest version will be used. |
p |
Platform string. |
pv |
Platform version string. |
m |
Architecture string. |
v
is the most tricky one:
- If omitted omnitruck will return information for the latest available version.
- When set to a full version string, omnitruck will return information for that specific version.
- When set to a partial string, omnitruck will return information for the latest version available that is matching the given string. E.g. if
"12"
is given, omnitruck will return the latest"12.X.Y"
and if"12.6"
is given, omnitruck will return the latest"12.6.Y"
.
The values that p
, pv
& m
can take include but are not limited to:
p
=> "aix", "debian", "el", "freebsd", "ios_xr", "mac_os_x", "nexus", "solaris2", "ubuntu", "windows", "solaris", "sles", "suse", "arista_eos"
pv
=> MAJOR.MINOR
version of the platform. E.g. "6.1", "7.1" for "aix", "6", "7", "8" for debian. For windows "2008r2"
m
=> "x86_64", "i386", "powerpc", "sparc", "ppc64le", "ppc64", "s390x"
Supports the same options with the /$channel/$project/metadata
endpoint and instead of returning package information, it returns a redirect to the "url"
set in the package information.
This api is similar to /$channel/$project/metadata
but instead of returning information about a single package, it returns information about all packages that have the same version number.
Response is an array of objects returned by /$channel/$project/metadata
endpoint. Supports v
but does not support p
, pv
or m
.
Returns a list of available version numbers for a particular channel and project combination
Returns the latest version number for a particular channel and project combination
Return the status of the application in json
format.
Historically omnitruck supported different naming conventions in its endpoints. So we have a big set deprecated endpoints. These endpoints are currently working and they are redirected to one of the supported endpoints.
- /download
- /metadata
- /download-server
- /metadata-server
- /full_client_list
- /full_list
- /full_server_list
- /chef/full_client_list
- /chef/full_list
- /chef/full_server_list
- /metadata-chefdk
- /download-chefdk
- /chef_platform_names
- /chef_server_platform_names
- /chef/chef_platform_names
- /chef/chef_server_platform_names
- /chef/metadata-chefdk
- /chef/download-chefdk
- /chef/metadata-container
- /chef/download-container
- /chef/metadata-angrychef
- /chef/download-angrychef
- /chef/download-server
- /chef/metadata-server
- /chef/install.msi
- /install.msi
- /full_$project_list
- /$project_platform_names
- /$channel/$project/platforms
- /$channel/$project/versions
./poller
is a tool that populates a Redis cache with information about the available versions of Chef Software, Inc. packages. In production is runs every 5 minutes with a cron job.
Version resolution is the most complex and fragile section of omnitruck. In addition to handling historical differences between the packages published by Chef (different package naming conventions, different metadata, different architecture naming) it supports some interesting logic to find the right package for a given platform.
PlatformDSL
and the main configuration file platforms.rb
gives omnitruck the ability to map builds released for a platform to be available for another platform. Some of the use cases enabled in platforms.rb
are:
- Making RHEL 6.4 and RHEL 6.5 be considered the same version but not considering Ubuntu 12.04 and Ubuntu 12.10 as same.
- Making centos, oracle, scientific, etc all map to the "el" platform.
The convention is to keep the platform names that install.sh supplies consistent with ohai platform names.
This convention was not always applied consistently in the past, so that "el" is the internal name for RHEL artifacts. The install.sh code has been changed so that "el" is no longer a distro name which install.sh detects (but omnitruck will still respond to "el" like it is a platform of "redhat").
All of the platform name and platform_version mangling should be moved from install.sh into this configuration file.
We only test certain artifacts running on certain distros and do not have complete coverage across all distros that can run omnibus artifacts. For example, we do no testing of Linux Mint at all. We also lag the distribution release, since getting a new tester into our CI testing is not as easy as it should be (particularly for distros like Mac OSX where we have laptops running jenkins as testers -- although this is getting better virtualized now). What "yolo" mode does is offer an artifact that we believe will work fine on the distribution, and will warn that user that we have not tested on that platform.
This will also let users deploy old versions of Chef on new distros where we never tested those old versions on that distro. The previous distro version will get promoted by yolo.
Packages which have been yolo-promoted have "yolo: true" in their JSON output from the metadata endpoint. The install.sh script prints a banner warning if it encounters a yolo mode package. It is up to the end user as to if they're willing to accept the risk of installing yolo-promoted packages (generally desktops and testing infrastructure is fine, but production infrastructure use is discouraged).
The shift in mac_os_x
from a MAJOR.MINOR versioning scheme (e.g., macOS 10.15) to a MAJOR-only versioning scheme (e.g., macOS 11) necessitated a modification in our version sorting logic. This is due to omnibus-related behavior that would identify, upload, and tag macOS 11 builds using MAJOR.MINOR versions (e.g., 11.0, 11.1, 11.2, etc) when the relevant platform version was only the MAJOR version (e.g., 11).
To handle environments that have a mix of equivalent MAJOR.MINOR and MAJOR builds in their artifact stores, we now compare versions using only shared version elements. This ensures that versions are sorted correctly, while still respecting Yolo mode behavior.
Example Comparison | Fallback Comparison | Result
--- | ---
"11" <=> "11.0"
| "11" <=> "11"
| Equivalent
"11.2" <=> "11"
| "11" <=> "11"
| Equivalent
"10.15" <=> "11"
| "10" <=> "11"
| Less Than
There are two parts to running omnitruck.
First, you need to populate its cache. Redis is used for the cache, and you will want to have a redis server running beforehand:
docker pull redis:7.4.0-alpine3.20
then to run it:
docker run --cpus 4 --memory 8g -it --privileged -p LOCAL_IP:6379:6379 IMAGE_NAME_OR_HASH
ex: docker run --cpus 4 --memory 8g -it --privileged -p 172.20.0.178:6379:6379 --volume $(pwd):/home 7de0dedd123b8cb2b105ace4e8d00b8bba5ad7be39617dfa229acff315fe4fbf
# Choose the following to install redis based on your OS
brew install redis
apt-get install redis
# Then just start it manually
redis-server
Then follow directions in Updating Mock Data
In production the cache populating is handled by a cron job. For development, you need to do this manually:
bundle install
bundle exec ./poller -e development
Omnitruck and the poller default to localhost when it looks for a redis
server, so you don't need to do anything special to get it to use the server
you started earlier. If you have redis running elsewhere, then you will want
to set the REDIS_URL
environment variable to point it at the correct server.
Second part is to run omnitruck web server. In production it runs in unicorn. For development, you have two options:
shotgun
does not output any of the logs, but reloads the application for each request so that you do not need to reload the application when you change a file.unicorn
gives you all the logs but requires you to restart when you change a file.
gem install shotgun
shotgun config.ru
bundle install
bundle exec unicorn
- Execute a special implementation of the poller
bundle install bundle exec rake refresh_data
- Update the latest version methods in
spec/spec_helper.rb
- Update any tests that may no longer be accurate. This is especially true for tests that expect a specific package or package version to exist in the current channel. Artifacts in the current channel expire after a certain time, so tests may become invalid.
- once all the data in the spec/data folder is downloaded (fyi this takes A LONG time if the file is out of date)
- load the data into your local running redis server by dropping to root and running
ruby redisManual.rb
- this is still a WIP, the format that the file is in now does load the data into the format that is expected.
- Copyright:: Copyright (c) 2010-2019 Chef Software, Inc.
- License:: Apache License, Version 2.0
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.