pylorax.api package¶
Submodules¶
pylorax.api.bisect module¶
-
pylorax.api.bisect.insort_left(a, x, key=None, lo=0, hi=None)[source]¶ Insert item x in list a, and keep it sorted assuming a is sorted.
- Parameters
- Returns
index where the item was inserted
- Return type
If x is already in a, insert it to the left of the leftmost x. Optional args lo (default 0) and hi (default len(a)) bound the slice of a to be searched.
This is a modified version of bisect.insort_left that can use a function for the compare, and returns the index position where it was inserted.
pylorax.api.cmdline module¶
pylorax.api.compose module¶
Setup for composing an image
Adding New Output Types¶
The new output type must add a kickstart template to ./share/composer/ where the name of the kickstart (without the trailing .ks) matches the entry in compose_args.
The kickstart should not have any url or repo entries, these will be added at build time. The %packages section should be the last thing, and while it can contain mandatory packages required by the output type, it should not have the trailing %end because the package NEVRAs will be appended to it at build time.
compose_args should have a name matching the kickstart, and it should set the novirt_install parameters needed to generate the desired output. Other types should be set to False.
-
pylorax.api.compose.add_customizations(f, recipe)[source]¶ Add customizations to the kickstart file
- Parameters
f (open file object) -- kickstart file object
recipe (Recipe object) --
- Returns
None
- Raises
RuntimeError if there was a problem writing to the kickstart
-
pylorax.api.compose.bootloader_append(line, kernel_append)[source]¶ Insert the kernel_append string into the --append argument
- Parameters
Using pykickstart to process the line is the best way to make sure it is parsed correctly, and re-assembled for inclusion into the final kickstart
-
pylorax.api.compose.compose_args(compose_type)[source]¶ Returns the settings to pass to novirt_install for the compose type
- Parameters
compose_type (str) -- The type of compose to create, from compose_types()
This will return a dict of options that match the ArgumentParser options for livemedia-creator. These are the ones the define the type of output, it's filename, etc. Other options will be filled in by make_compose()
-
pylorax.api.compose.compose_types(share_dir)[source]¶ Returns a list of the supported output types
The output types come from the kickstart names in /usr/share/lorax/composer/*ks
-
pylorax.api.compose.customize_ks_template(ks_template, recipe)[source]¶ Customize the kickstart template and return it
- Parameters
ks_template (str) -- The kickstart template
recipe (Recipe object) --
Apply customizations.kernel.append to the bootloader argument in the template. Add bootloader line if it is missing.
-
pylorax.api.compose.get_extra_pkgs(dbo, share_dir, compose_type)[source]¶ Return extra packages needed for the output type
- Parameters
- Returns
List of package names (name only, not NEVRA)
- Return type
Currently this is only needed by live-iso, it reads ./live/live-install.tmpl and processes only the installpkg lines. It lists the packages needed to complete creation of the iso using the templates such as x86.tmpl
Keep in mind that the live-install.tmpl is shared between livemedia-creator and lorax-composer, even though the results are applied differently.
-
pylorax.api.compose.get_kernel_append(recipe)[source]¶ Return the customizations.kernel append value
- Parameters
recipe (Recipe object) --
- Returns
append value or empty string
- Return type
-
pylorax.api.compose.move_compose_results(cfg, results_dir)[source]¶ Move the final image to the results_dir and cleanup the unneeded compose files
- Parameters
cfg (DataHolder) -- Build configuration
results_dir (str) -- Directory to put the results into
-
pylorax.api.compose.repo_to_ks(r, url='url')[source]¶ Return a kickstart line with the correct args. :param r: DNF repository information :type r: dnf.Repo :param url: "url" or "baseurl" to use for the baseurl parameter :type url: str :returns: kickstart command arguments for url/repo command :rtype: str
Set url to "baseurl" if it is a repo, leave it as "url" for the installation url.
-
pylorax.api.compose.start_build(cfg, dnflock, gitlock, branch, recipe_name, compose_type, test_mode=0)[source]¶ Start the build
- Parameters
cfg (ComposerConfig) -- Configuration object
dnflock (YumLock) -- Lock and YumBase for depsolving
recipe (str) -- The recipe to build
compose_type (str) -- The type of output to create from the recipe
- Returns
Unique ID for the build that can be used to track its status
- Return type
-
pylorax.api.compose.test_templates(dbo, share_dir)[source]¶ Try depsolving each of the the templates and report any errors
- Parameters
dbo (dnf.Base) -- dnf base object
- Returns
List of template types and errors
- Return type
List of errors
Return a list of templates and errors encountered or an empty list
-
pylorax.api.compose.write_ks_group(f, group)[source]¶ Write kickstart group entry
- Parameters
f (open file object) -- kickstart file object
group -- A blueprint group dictionary
gid is optional
-
pylorax.api.compose.write_ks_root(f, user)[source]¶ Write kickstart root password and sshkey entry
- Parameters
f (open file object) -- kickstart file object
user (dict) -- A blueprint user dictionary
- Returns
True if it wrote a rootpw command to the kickstart
- Return type
If the entry contains a ssh key, use sshkey to write it If it contains password, use rootpw to set it
root cannot be used with the user command. So only key and password are supported for root.
-
pylorax.api.compose.write_ks_user(f, user)[source]¶ Write kickstart user and sshkey entry
- Parameters
f (open file object) -- kickstart file object
user (dict) -- A blueprint user dictionary
If the entry contains a ssh key, use sshkey to write it All of the user fields are optional, except name, write out a kickstart user entry with whatever options are relevant.
pylorax.api.config module¶
-
class
pylorax.api.config.ComposerConfig(defaults=None, dict_type=<class 'collections.OrderedDict'>, allow_no_value=False, *, delimiters=('=', ':'), comment_prefixes=('#', ';'), inline_comment_prefixes=None, strict=True, empty_lines_in_values=True, default_section='DEFAULT', interpolation=<object object>, converters=<object object>)[source]¶ Bases:
configparser.ConfigParser
-
pylorax.api.config.configure(conf_file='/etc/lorax/composer.conf', root_dir='/', test_config=False)[source]¶ lorax-composer configuration
-
pylorax.api.config.make_dnf_dirs(conf, uid, gid)[source]¶ Make any missing dnf directories owned by user:group
- Parameters
conf (ComposerConfig) -- The configuration to use
uid (int) -- uid of owner
gid (int) -- gid of owner
- Returns
list of errors
- Return type
list of str
-
pylorax.api.config.make_owned_dir(p_dir, uid, gid)[source]¶ Make a directory and its parents, setting owner and group
- Parameters
- Returns
list of errors
- Return type
list of str
Check to make sure it does not have o+rw permissions and that it is owned by uid:gid
-
pylorax.api.config.make_queue_dirs(conf, gid)[source]¶ Make any missing queue directories
- Parameters
conf (ComposerConfig) -- The configuration to use
gid (int) -- Group ID that has access to the queue directories
- Returns
list of errors
- Return type
list of str
pylorax.api.crossdomain module¶
pylorax.api.dnfbase module¶
-
class
pylorax.api.dnfbase.DNFLock(conf, expire_secs=21600)[source]¶ Bases:
objectHold the dnf.Base object and a Lock to control access to it.
self.dbo is a property that returns the dnf.Base object, but it may change from one call to the next if the upstream repositories have changed.
-
lock¶ Check for repo updates (using expiration time) and return the lock
If the repository has been updated, tear down the old dnf.Base and create a new one. This is the only way to force dnf to use the new metadata.
-
lock_check¶ Force a check for repo updates and return the lock
Use this method sparingly, it removes the repodata and downloads a new copy every time.
-
pylorax.api.errors module¶
pylorax.api.gitrpm module¶
Clone a git repository and package it as an rpm
This module contains functions for cloning a git repo, creating a tar archive of the selected commit, branch, or tag, and packaging the files into an rpm that will be installed by anaconda when creating the image.
-
class
pylorax.api.gitrpm.GitArchiveTarball(gitRepo)[source]¶ Bases:
objectCreate a git archive of the selected git repo and reference
-
class
pylorax.api.gitrpm.GitRpmBuild(*args, **kwargs)[source]¶ Bases:
rpmfluff.SimpleRpmBuildBuild an rpm containing files from a git repository
-
add_git_tarball(gitRepo)[source]¶ Add a tar archive of a git repository to the rpm
- Parameters
gitRepo (dict) -- A dict with the repository details
This populates the rpm with the URL of the git repository, the summary describing the repo, the description of the repository and reference used, and sets up the rpm to install the archive contents into the destination path.
-
-
pylorax.api.gitrpm.create_gitrpm_repo(results_dir, recipe)[source]¶ Create a dnf repository with the rpms from the recipe
- Parameters
- Returns
Path to the dnf repository or ""
- Return type
This function creates a dnf repository directory at results_dir+"repo/", creates rpms for all of the repos.git entries in the recipe, runs createrepo_c on the dnf repository so that Anaconda can use it, and returns the path to the repository to the caller.
-
pylorax.api.gitrpm.get_repo_description(gitRepo)[source]¶ Return a description including the git repo and reference
-
pylorax.api.gitrpm.make_git_rpm(gitRepo, dest)[source]¶ Create an rpm from the specified git repo
- Parameters
gitRepo (dict) -- A dict with the repository details
This will clone the git repository, create an archive of the selected reference, and build an rpm that will install the files from the repository under the destination directory. The gitRepo dict should have the following fields:
rpmname: "server-config" rpmversion: "1.0" rpmrelease: "1" summary: "Setup files for server deployment" repo: "PATH OF GIT REPO TO CLONE" ref: "v1.0" destination: "/opt/server/"
rpmname: Name of the rpm to create, also used as the prefix name in the tar archive
rpmversion: Version of the rpm, eg. "1.0.0"
rpmrelease: Release of the rpm, eg. "1"
summary: Summary string for the rpm
repo: URL of the get repo to clone and create the archive from
ref: Git reference to check out. eg. origin/branch-name, git tag, or git commit hash
destination: Path to install the / of the git repo at when installing the rpm
pylorax.api.projects module¶
-
pylorax.api.projects.api_changelog(changelog)[source]¶ Convert the changelog to a string
- Parameters
changelog (tuple) -- A list of time, author, string tuples.
- Returns
The most recent changelog text or ""
- Return type
This returns only the most recent changelog entry.
-
pylorax.api.projects.delete_repo_source(source_glob, source_name)[source]¶ Delete a source from a repo file
- Parameters
source_glob (str) -- A glob of the repo sources to search
- Returns
None
- Raises
ProjectsError if there was a problem
A repo file may have multiple sources in it, delete only the selected source. If it is the last one in the file, delete the file.
WARNING: This will delete ANY source, the caller needs to ensure that a system source_name isn't passed to it.
-
pylorax.api.projects.dnf_repo_to_file_repo(repo)[source]¶ Return a string representation of a DNF Repo object suitable for writing to a .repo file
- Parameters
repo (dnf.RepoDict) -- DNF Repository
- Returns
A string
- Return type
The DNF Repo.dump() function does not produce a string that can be used as a dnf .repo file, it ouputs baseurl and gpgkey as python lists which DNF cannot read. So do this manually with only the attributes we care about.
-
pylorax.api.projects.estimate_size(packages, block_size=6144)[source]¶ Estimate the installed size of a package list
- Parameters
packages (list of hawkey.Package objects) -- The packages to be installed
block_size (int) -- The block size to use for rounding up file sizes.
- Returns
The estimated size of installed packages
- Return type
Estimating actual requirements is difficult without the actual file sizes, which dnf doesn't provide access to. So use the file count and block size to estimate a minimum size for each package.
-
pylorax.api.projects.get_repo_sources(source_glob)[source]¶ Return a list of sources from a directory of yum repositories
- Parameters
source_glob (str) -- A glob to use to match the source files, including full path
- Returns
A list of the source ids in all of the matching files
- Return type
list of str
-
pylorax.api.projects.get_source_ids(source_path)[source]¶ Return a list of the source ids in a file
- Parameters
source_path (str) -- Full path and filename of the source (yum repo) file
- Returns
A list of source id strings
- Return type
list of str
-
pylorax.api.projects.modules_info(dbo, module_names)[source]¶ Return details about a module, including dependencies
- Parameters
dbo (dnf.Base) -- dnf base object
module_names (str) -- Names of the modules to get info about
- Returns
List of dicts with module details and dependencies.
- Return type
list of dicts
-
pylorax.api.projects.modules_list(dbo, module_names)[source]¶ Return a list of modules
- Parameters
dbo (dnf.Base) -- dnf base object
offset -- Number of modules to skip
limit (int) -- Maximum number of modules to return
- Returns
List of module information and total count
- Return type
tuple of a list of dicts and an Int
Modules don't exist in RHEL7 so this only returns projects and sets the type to "rpm"
-
pylorax.api.projects.pkg_to_build(pkg)[source]¶ Extract the build details from a hawkey.Package object
- Parameters
pkg (hawkey.Package) -- hawkey.Package object with package details
- Returns
A dict with the build details, epoch, release, arch, build_time, changelog, ...
- Return type
metadata entries are hard-coded to {}
Note that this only returns the build dict, it does not include the name, description, etc.
-
pylorax.api.projects.pkg_to_dep(pkg)[source]¶ Extract the info from a hawkey.Package object
- Parameters
pkg (hawkey.Package) -- A hawkey.Package object
- Returns
A dict with name, epoch, version, release, arch
- Return type
-
pylorax.api.projects.pkg_to_project(pkg)[source]¶ Extract the details from a hawkey.Package object
- Parameters
pkgs (hawkey.Package) -- hawkey.Package object with package details
- Returns
A dict with the name, summary, description, and url.
- Return type
upstream_vcs is hard-coded to UPSTREAM_VCS
-
pylorax.api.projects.pkg_to_project_info(pkg)[source]¶ Extract the details from a hawkey.Package object
- Parameters
pkg (hawkey.Package) -- hawkey.Package object with package details
- Returns
A dict with the project details, as well as epoch, release, arch, build_time, changelog, ...
- Return type
metadata entries are hard-coded to {}
-
pylorax.api.projects.proj_to_module(proj)[source]¶ Extract the name from a project_info dict
- Parameters
pkg (dict) -- dict with package details
- Returns
A dict with name, and group_type
- Return type
group_type is hard-coded to "rpm"
-
pylorax.api.projects.projects_depsolve(dbo, projects, groups)[source]¶ Return the dependencies for a list of projects
- Parameters
dbo (dnf.Base) -- dnf base object
projects (List of Strings) -- The projects to find the dependencies for
groups (List of str) -- The groups to include in dependency solving
- Returns
NEVRA's of the project and its dependencies
- Return type
list of dicts
- Raises
ProjectsError if there was a problem installing something
-
pylorax.api.projects.projects_depsolve_with_size(dbo, projects, groups, with_core=True)[source]¶ Return the dependencies and installed size for a list of projects
- Parameters
dbo (dnf.Base) -- dnf base object
project_names (List of Strings) -- The projects to find the dependencies for
groups (List of str) -- The groups to include in dependency solving
- Returns
installed size and a list of NEVRA's of the project and its dependencies
- Return type
tuple of (int, list of dicts)
- Raises
ProjectsError if there was a problem installing something
-
pylorax.api.projects.projects_info(dbo, project_names)[source]¶ Return details about specific projects
- Parameters
dbo (dnf.Base) -- dnf base object
project_names (str) -- List of names of projects to get info about
- Returns
List of project info dicts with pkg_to_project as well as epoch, version, release, etc.
- Return type
list of dicts
If project_names is None it will return the full list of available packages
-
pylorax.api.projects.projects_list(dbo)[source]¶ Return a list of projects
- Parameters
dbo (dnf.Base) -- dnf base object
- Returns
List of project info dicts with name, summary, description, homepage, upstream_vcs
- Return type
list of dicts
-
pylorax.api.projects.repo_to_source(repo, system_source)[source]¶ Return a Weldr Source dict created from the DNF Repository
- Parameters
repo (dnf.RepoDict) -- DNF Repository
system_source (bool) -- True if this source is an immutable system source
- Returns
A dict with Weldr Source fields filled in
- Return type
Example:
{ "check_gpg": true, "check_ssl": true, "gpgkey_url": [ "file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora-28-x86_64" ], "name": "fedora", "proxy": "http://proxy.brianlane.com:8123", "system": true "type": "yum-metalink", "url": "https://mirrors.fedoraproject.org/metalink?repo=fedora-28&arch=x86_64" }
-
pylorax.api.projects.source_to_repo(source, dnf_conf)[source]¶ Return a dnf Repo object created from a source dict
- Parameters
source (dict) -- A Weldr source dict
- Returns
A dnf Repo object
- Return type
dnf.Repo
Example:
{ "check_gpg": True, "check_ssl": True, "gpgkey_urls": [ "file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora-28-x86_64" ], "name": "fedora", "proxy": "http://proxy.brianlane.com:8123", "system": True "type": "yum-metalink", "url": "https://mirrors.fedoraproject.org/metalink?repo=fedora-28&arch=x86_64" }
pylorax.api.queue module¶
Functions to monitor compose queue and run anaconda
-
pylorax.api.queue.build_status(cfg, status_filter=None)[source]¶ Return the details of finished or failed builds
- Parameters
cfg (ComposerConfig) -- Configuration settings
status_filter (str) -- What builds to return. None == all, "FINISHED", or "FAILED"
- Returns
A list of the build details (from compose_details)
- Return type
list of dicts
This returns a list of build details for each of the matching builds on the system. It does not return the status of builds that have not been finished. Use queue_status() for those.
-
pylorax.api.queue.check_queues(cfg)[source]¶ Check to make sure the new and run queue symlinks are correct
- Parameters
cfg (DataHolder) -- Configuration settings
Also check all of the existing results and make sure any with WAITING set in STATUS have a symlink in queue/new/
-
pylorax.api.queue.compose_detail(results_dir)[source]¶ Return details about the build.
- Parameters
results_dir (str) -- The directory containing the metadata and results for the build
- Returns
A dictionary with details about the compose
- Return type
- Raises
IOError if it cannot read the directory, STATUS, or blueprint file.
The following details are included in the dict:
id - The uuid of the comoposition
queue_status - The final status of the composition (FINISHED or FAILED)
compose_type - The type of output generated (tar, iso, etc.)
blueprint - Blueprint name
version - Blueprint version
image_size - Size of the image, if finished. 0 otherwise.
Various timestamps are also included in the dict. These are all Unix UTC timestamps. It is possible for these timestamps to not always exist, in which case they will be None in Python (or null in JSON). The following timestamps are included:
job_created - When the user submitted the compose
job_started - Anaconda started running
job_finished - Job entered FINISHED or FAILED state
-
pylorax.api.queue.get_image_name(uuid_dir)[source]¶ Return the filename and full path of the build's image file
- Parameters
uuid (str) -- The UUID of the build
- Returns
The image filename and full path
- Return type
tuple of strings
- Raises
RuntimeError if there was a problem (eg. invalid uuid, missing config file)
-
pylorax.api.queue.make_compose(cfg, results_dir)[source]¶ Run anaconda with the final-kickstart.ks from results_dir
- Parameters
cfg (DataHolder) -- Configuration settings
results_dir (str) -- The directory containing the metadata and results for the build
- Returns
Nothing
- Raises
May raise various exceptions
This takes the final-kickstart.ks, and the settings in config.toml and runs Anaconda in no-virt mode (directly on the host operating system). Exceptions should be caught at the higer level.
If there is a failure, the build artifacts will be cleaned up, and any logs will be moved into logs/anaconda/ and their ownership will be set to the user from the cfg object.
-
pylorax.api.queue.monitor(cfg)[source]¶ Monitor the queue for new compose requests
- Parameters
cfg (DataHolder) -- Configuration settings
- Returns
Does not return
The queue has 2 subdirectories, new and run. When a compose is ready to be run a symlink to the uniquely named results directory should be placed in ./queue/new/
When the it is ready to be run (it is checked every 30 seconds or after a previous compose is finished) the symlink will be moved into ./queue/run/ and a STATUS file will be created in the results directory.
STATUS can contain one of: WAITING, RUNNING, FINISHED, FAILED
If the system is restarted while a compose is running it will move any old symlinks from ./queue/run/ to ./queue/new/ and rerun them.
-
pylorax.api.queue.queue_status(cfg)[source]¶ Return details about what is in the queue.
- Parameters
cfg (ComposerConfig) -- Configuration settings
- Returns
A list of the new composes, and a list of the running composes
- Return type
This returns a dict with 2 lists. "new" is the list of uuids that are waiting to be built, and "run" has the uuids that are being built (currently limited to 1 at a time).
-
pylorax.api.queue.start_queue_monitor(cfg, uid, gid)[source]¶ Start the queue monitor as a mp process
- Parameters
cfg (ComposerConfig) -- Configuration settings
uid (int) -- User ID that owns the queue
gid (int) -- Group ID that owns the queue
- Returns
None
-
pylorax.api.queue.uuid_cancel(cfg, uuid)[source]¶ Cancel a build and delete its results
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
- Returns
True if it was canceled and deleted
- Return type
Only call this if the build status is WAITING or RUNNING
-
pylorax.api.queue.uuid_delete(cfg, uuid)[source]¶ Delete all of the results from a compose
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
- Returns
True if it was deleted
- Return type
- Raises
This will raise an error if the delete failed
-
pylorax.api.queue.uuid_image(cfg, uuid)[source]¶ Return the filename and full path of the build's image file
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
- Returns
The image filename and full path
- Return type
tuple of strings
- Raises
RuntimeError if there was a problem (eg. invalid uuid, missing config file)
-
pylorax.api.queue.uuid_info(cfg, uuid)[source]¶ Return information about the composition
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
- Returns
dictionary of information about the composition or None
- Return type
- Raises
RuntimeError if there was a problem
This will return a dict with the following fields populated:
id - The uuid of the comoposition
config - containing the configuration settings used to run Anaconda
blueprint - The depsolved blueprint used to generate the kickstart
commit - The (local) git commit hash for the blueprint used
deps - The NEVRA of all of the dependencies used in the composition
compose_type - The type of output generated (tar, iso, etc.)
queue_status - The final status of the composition (FINISHED or FAILED)
-
pylorax.api.queue.uuid_log(cfg, uuid, size=1024)[source]¶ Return size kbytes from the end of the anaconda.log
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
size (int) -- Number of kbytes to read. Default is 1024
- Returns
Up to size kbytes from the end of the log
- Return type
- Raises
RuntimeError if there was a problem (eg. no log file available)
This function tries to return lines from the end of the log, it will attempt to start on a line boundry, and may return less than size kbytes.
-
pylorax.api.queue.uuid_status(cfg, uuid)[source]¶ Return the details of a specific UUID compose
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
- Returns
Details about the build
- Return type
Returns the same dict as compose_details()
-
pylorax.api.queue.uuid_tar(cfg, uuid, metadata=False, image=False, logs=False)[source]¶ Return a tar of the build data
- Parameters
cfg (ComposerConfig) -- Configuration settings
uuid (str) -- The UUID of the build
metadata (bool) -- Set to true to include all the metadata needed to reproduce the build
image (bool) -- Set to true to include the output image
logs (bool) -- Set to true to include the logs from the build
- Returns
A stream of bytes from tar
- Return type
A generator
- Raises
RuntimeError if there was a problem (eg. missing config file)
This yields an uncompressed tar's data to the caller. It includes the selected data to the caller by returning the Popen stdout from the tar process.
pylorax.api.recipes module¶
-
class
pylorax.api.recipes.CommitDetails(commit, timestamp, message, revision=None)[source]¶ Bases:
pylorax.base.DataHolder
-
pylorax.api.recipes.NewRecipeGit(toml_dict)[source]¶ Create a RecipeGit object from fields in a TOML dict
- Parameters
rpmname (str) -- Name of the rpm to create, also used as the prefix name in the tar archive
rpmversion (str) -- Version of the rpm, eg. "1.0.0"
rpmrelease (str) -- Release of the rpm, eg. "1"
summary (str) -- Summary string for the rpm
repo (str) -- URL of the get repo to clone and create the archive from
ref (str) -- Git reference to check out. eg. origin/branch-name, git tag, or git commit hash
destination (str) -- Path to install the / of the git repo at when installing the rpm
- Returns
A populated RecipeGit object
- Return type
The TOML should look like this:
[[repos.git]] rpmname="server-config" rpmversion="1.0" rpmrelease="1" summary="Setup files for server deployment" repo="PATH OF GIT REPO TO CLONE" ref="v1.0" destination="/opt/server/"
Note that the repo path supports anything that git supports, file://, https://, http://
Currently there is no support for authentication
-
class
pylorax.api.recipes.Recipe(name, description, version, modules, packages, groups, customizations=None, gitrepos=None)[source]¶ Bases:
dictA Recipe of package and modules
This is a subclass of dict that enforces the constructor arguments and adds a .filename property to return the recipe's filename, and a .toml() function to return the recipe as a TOML string.
-
bump_version(old_version=None)[source]¶ semver recipe version number bump
- Parameters
old_version (str) -- An optional old version number
- Returns
The new version number or None
- Return type
- Raises
ValueError
If neither have a version, 0.0.1 is returned If there is no old version the new version is checked and returned If there is no new version, but there is a old one, bump its patch level If the old and new versions are the same, bump the patch level If they are different, check and return the new version
-
filename¶ Return the Recipe's filename
Replaces spaces in the name with '-' and appends .toml
-
group_names¶ Return the names of the groups. Groups do not have versions.
-
module_names¶ Return the names of the modules
-
module_nver¶ Return the names and version globs of the modules
-
package_names¶ Return the names of the packages
-
package_nver¶ Return the names and version globs of the packages
-
-
class
pylorax.api.recipes.RecipeGit(rpmname, rpmversion, rpmrelease, summary, repo, ref, destination)[source]¶ Bases:
dict
-
pylorax.api.recipes.commit_recipe_directory(repo, branch, directory)[source]¶ Commit all *.toml files from a directory, if they aren't already in git.
- Parameters
- Returns
None
- Raises
Can raise errors from Ggit or RecipeFileError
Files with Toml or RecipeFileErrors will be skipped, and the remainder will be tried.
-
pylorax.api.recipes.commit_recipe_file(repo, branch, filename)[source]¶ Commit a recipe file to a branch
-
pylorax.api.recipes.delete_recipe(repo, branch, recipe_name)[source]¶ Delete a recipe from a branch.
-
pylorax.api.recipes.diff_items(title, old_items, new_items)[source]¶ Return the differences between two lists of dicts.
-
pylorax.api.recipes.find_commit_tag(repo, branch, filename, commit_id)[source]¶ Find the tag that matches the commit_id
- Parameters
- Returns
The tag or None if there isn't one
- Return type
There should be only 1 tag pointing to a commit, but there may not be a tag at all.
The tag will look like: 'refs/tags/<branch>/<filename>/r<revision>'
-
pylorax.api.recipes.find_name(name, lst)[source]¶ Find the dict matching the name in a list and return it.
-
pylorax.api.recipes.get_commit_details(commit, revision=None)[source]¶ Return the details about a specific commit.
- Parameters
commit (Git.Commit) -- The commit to get details from
revision (int) -- Optional commit revision
- Returns
Details about the commit
- Return type
- Raises
CommitTimeValError or Ggit exceptions
-
pylorax.api.recipes.get_revision_from_tag(tag)[source]¶ Return the revision number from a tag
- Parameters
tag (str) -- The tag to exract the revision from
- Returns
The integer revision or None
- Return type
The revision is the part after the r in 'branch/filename/rXXX'
-
pylorax.api.recipes.head_commit(repo, branch)[source]¶ Get the branch's HEAD Commit Object
- Parameters
repo (Git.Repository) -- Open repository
branch (str) -- Branch name
- Returns
Branch's head commit
- Return type
Git.Commit
- Raises
Can raise errors from Ggit
-
pylorax.api.recipes.is_commit_tag(repo, commit_id, tag)[source]¶ Check to see if a tag points to a specific commit.
-
pylorax.api.recipes.is_parent_diff(repo, filename, tree, parent)[source]¶ Check to see if the commit is different from its parents
-
pylorax.api.recipes.list_branch_files(repo, branch)[source]¶ Return a sorted list of the files on the branch HEAD
-
pylorax.api.recipes.list_commit_files(repo, commit)[source]¶ Return a sorted list of the files on a commit
-
pylorax.api.recipes.list_commits(repo, branch, filename, limit=0)[source]¶ List the commit history of a file on a branch.
- Parameters
- Returns
A list of commit details
- Return type
- Raises
Can raise errors from Ggit
-
pylorax.api.recipes.open_or_create_repo(path)[source]¶ Open an existing repo, or create a new one
- Parameters
path (string) -- path to recipe directory
- Returns
A repository object
- Return type
Git.Repository
- Raises
Can raise errors from Ggit
A bare git repo will be created in the git directory of the specified path. If a repo already exists it will be opened and returned instead of creating a new one.
-
pylorax.api.recipes.prepare_commit(repo, branch, builder)[source]¶ Prepare for a commit
- Parameters
repo (Git.Repository) -- Open repository
branch (str) -- Branch name
builder (TreeBuilder) -- instance of TreeBuilder
- Returns
(Tree, Sig, Ref)
- Return type
- Raises
Can raise errors from Ggit
-
pylorax.api.recipes.read_commit(repo, branch, filename, commit=None)[source]¶ Return the contents of a file on a specific branch or commit.
- Parameters
- Returns
The commit id, and the contents of the commit
- Return type
- Raises
Can raise errors from Ggit
If no commit is passed the master:filename is returned, otherwise it will be commit:filename
-
pylorax.api.recipes.read_commit_spec(repo, spec)[source]¶ Return the raw content of the blob specified by the spec
- Parameters
repo (Git.Repository) -- Open repository
spec (str) -- Git revparse spec
- Returns
Contents of the commit
- Return type
- Raises
Can raise errors from Ggit
eg. To read the README file from master the spec is "master:README"
-
pylorax.api.recipes.read_recipe_and_id(repo, branch, recipe_name, commit=None)[source]¶ Read a recipe commit and its id from git
- Parameters
- Returns
The commit id, and a Recipe object
- Return type
- Raises
Can raise errors from Ggit
If no commit is passed the master:filename is returned, otherwise it will be commit:filename
-
pylorax.api.recipes.read_recipe_commit(repo, branch, recipe_name, commit=None)[source]¶ Read a recipe commit from git and return a Recipe object
- Parameters
- Returns
A Recipe object
- Return type
- Raises
Can raise errors from Ggit
If no commit is passed the master:filename is returned, otherwise it will be commit:filename
-
pylorax.api.recipes.recipe_filename(name)[source]¶ Return the toml filename for a recipe
Replaces spaces with '-' and appends '.toml'
-
pylorax.api.recipes.recipe_from_dict(recipe_dict)[source]¶ Create a Recipe object from a plain dict.
-
pylorax.api.recipes.recipe_from_toml(recipe_str)[source]¶ Create a Recipe object from a toml string.
-
pylorax.api.recipes.repo_file_exists(repo, branch, filename)[source]¶ Return True if the filename exists on the branch
-
pylorax.api.recipes.revert_file(repo, branch, filename, commit)[source]¶ Revert the contents of a file to that of a previous commit
-
pylorax.api.recipes.revert_recipe(repo, branch, recipe_name, commit)[source]¶ Revert the contents of a recipe to that of a previous commit
-
pylorax.api.recipes.tag_file_commit(repo, branch, filename)[source]¶ Tag a file's most recent commit
- Parameters
- Returns
Tag id or None if it failed.
- Return type
Git.OId
- Raises
Can raise errors from Ggit
This uses git tags, of the form refs/tags/<branch>/<filename>/r<revision> Only the most recent recipe commit can be tagged to prevent out of order tagging. Revisions start at 1 and increment for each new commit that is tagged. If the commit has already been tagged it will return false.
-
pylorax.api.recipes.tag_recipe_commit(repo, branch, recipe_name)[source]¶ Tag a file's most recent commit
- Parameters
- Returns
Tag id or None if it failed.
- Return type
Git.OId
- Raises
Can raise errors from Ggit
Uses tag_file_commit()
pylorax.api.regexes module¶
pylorax.api.server module¶
pylorax.api.timestamp module¶
pylorax.api.v0 module¶
Setup v0 of the API server
v0_api() must be called to setup the API routes for Flask
Status Responses¶
Some requests only return a status/error response.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
Example response:
{ "status": true }Error response:
{ "errors": ["ggit-error: Failed to remove entry. File isn't in the tree - jboss.toml (-1)"] "status": false }
API Routes¶
All of the blueprints routes support the optional branch argument. If it is not used then the API will use the master branch for blueprints. If you want to create a new branch use the new or workspace routes with ?branch=<branch-name> to store the new blueprint on the new branch.
/api/v0/blueprints/list¶
List the available blueprints:
{ "limit": 20, "offset": 0, "blueprints": [ "atlas", "development", "glusterfs", "http-server", "jboss", "kubernetes" ], "total": 6 }
/api/v0/blueprints/info/<blueprint_names>[?format=<json|toml>]¶
Return the JSON representation of the blueprint. This includes 3 top level objects. changes which lists whether or not the workspace is different from the most recent commit. blueprints which lists the JSON representation of the blueprint, and errors which will list any errors, like non-existant blueprints.
By default the response is JSON, but if ?format=toml is included in the URL's arguments it will return the response as the blueprint's raw TOML content. Unless there is an error which will only return a 400 and a standard error `Status Response`_.
If there is an error when JSON is requested the successful blueprints and the errors will both be returned.
Example of json response:
{ "changes": [ { "changed": false, "name": "glusterfs" } ], "errors": [], "blueprints": [ { "description": "An example GlusterFS server with samba", "modules": [ { "name": "glusterfs", "version": "3.7.*" }, { "name": "glusterfs-cli", "version": "3.7.*" } ], "name": "glusterfs", "packages": [ { "name": "2ping", "version": "3.2.1" }, { "name": "samba", "version": "4.2.*" } ], "version": "0.0.6" } ] }Error example:
{ "changes": [], "errors": ["ggit-error: the path 'missing.toml' does not exist in the given tree (-3)"] "blueprints": [] }
/api/v0/blueprints/changes/<blueprint_names>[?offset=0&limit=20]¶
Return the commits to a blueprint. By default it returns the first 20 commits, this can be changed by passing offset and/or limit. The response will include the commit hash, summary, timestamp, and optionally the revision number. The commit hash can be passed to /api/v0/blueprints/diff/ to retrieve the exact changes.
Example:
{ "errors": [], "limit": 20, "offset": 0, "blueprints": [ { "changes": [ { "commit": "e083921a7ed1cf2eec91ad12b9ad1e70ef3470be", "message": "blueprint glusterfs, version 0.0.6 saved.", "revision": null, "timestamp": "2017-11-23T00:18:13Z" }, { "commit": "cee5f4c20fc33ea4d54bfecf56f4ad41ad15f4f3", "message": "blueprint glusterfs, version 0.0.5 saved.", "revision": null, "timestamp": "2017-11-11T01:00:28Z" }, { "commit": "29b492f26ed35d80800b536623bafc51e2f0eff2", "message": "blueprint glusterfs, version 0.0.4 saved.", "revision": null, "timestamp": "2017-11-11T00:28:30Z" }, { "commit": "03374adbf080fe34f5c6c29f2e49cc2b86958bf2", "message": "blueprint glusterfs, version 0.0.3 saved.", "revision": null, "timestamp": "2017-11-10T23:15:52Z" }, { "commit": "0e08ecbb708675bfabc82952599a1712a843779d", "message": "blueprint glusterfs, version 0.0.2 saved.", "revision": null, "timestamp": "2017-11-10T23:14:56Z" }, { "commit": "3e11eb87a63d289662cba4b1804a0947a6843379", "message": "blueprint glusterfs, version 0.0.1 saved.", "revision": null, "timestamp": "2017-11-08T00:02:47Z" } ], "name": "glusterfs", "total": 6 } ] }
POST /api/v0/blueprints/new¶
Create a new blueprint, or update an existing blueprint. This supports both JSON and TOML for the blueprint format. The blueprint should be in the body of the request with the Content-Type header set to either application/json or text/x-toml.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
DELETE /api/v0/blueprints/delete/<blueprint_name>¶
Delete a blueprint. The blueprint is deleted from the branch, and will no longer be listed by the list route. A blueprint can be undeleted using the undo route to revert to a previous commit. This will also delete the workspace copy of the blueprint.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
POST /api/v0/blueprints/workspace¶
Write a blueprint to the temporary workspace. This works exactly the same as new except that it does not create a commit. JSON and TOML bodies are supported.
The workspace is meant to be used as a temporary blueprint storage for clients. It will be read by the info and diff routes if it is different from the most recent commit.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
DELETE /api/v0/blueprints/workspace/<blueprint_name>¶
Remove the temporary workspace copy of a blueprint. The info route will now return the most recent commit of the blueprint. Any changes that were in the workspace will be lost.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
POST /api/v0/blueprints/undo/<blueprint_name>/<commit>¶
This will revert the blueprint to a previous commit. The commit hash from the changes route can be used in this request.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
POST /api/v0/blueprints/tag/<blueprint_name>¶
Tag a blueprint as a new release. This uses git tags with a special format. refs/tags/<branch>/<filename>/r<revision>. Only the most recent blueprint commit can be tagged. Revisions start at 1 and increment for each new tag (per-blueprint). If the commit has already been tagged it will return false.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
/api/v0/blueprints/diff/<blueprint_name>/<from_commit>/<to_commit>¶
Return the differences between two commits, or the workspace. The commit hash from the changes response can be used here, or several special strings:
NEWEST will select the newest git commit. This works for from_commit or to_commit
WORKSPACE will select the workspace copy. This can only be used in to_commit
eg. /api/v0/blueprints/diff/glusterfs/NEWEST/WORKSPACE will return the differences between the most recent git commit and the contents of the workspace.
Each entry in the response's diff object contains the old blueprint value and the new one. If old is null and new is set, then it was added. If new is null and old is set, then it was removed. If both are set, then it was changed.
The old/new entries will have the name of the blueprint field that was changed. This can be one of: Name, Description, Version, Module, or Package. The contents for these will be the old/new values for them.
In the example below the version was changed and the ping package was added.
Example:
{ "diff": [ { "new": { "Version": "0.0.6" }, "old": { "Version": "0.0.5" } }, { "new": { "Package": { "name": "ping", "version": "3.2.1" } }, "old": null } ] }
/api/v0/blueprints/freeze/<blueprint_names>¶
Return a JSON representation of the blueprint with the package and module versions set to the exact versions chosen by depsolving the blueprint.
Example:
{ "errors": [], "blueprints": [ { "blueprint": { "description": "An example GlusterFS server with samba", "modules": [ { "name": "glusterfs", "version": "3.8.4-18.4.el7.x86_64" }, { "name": "glusterfs-cli", "version": "3.8.4-18.4.el7.x86_64" } ], "name": "glusterfs", "packages": [ { "name": "ping", "version": "2:3.2.1-2.el7.noarch" }, { "name": "samba", "version": "4.6.2-8.el7.x86_64" } ], "version": "0.0.6" } } ] }
/api/v0/blueprints/depsolve/<blueprint_names>¶
Depsolve the blueprint using yum, return the blueprint used, and the NEVRAs of the packages chosen to satisfy the blueprint's requirements. The response will include a list of results, with the full dependency list in dependencies, the NEVRAs for the blueprint's direct modules and packages in modules, and any error will be in errors.
Example:
{ "errors": [], "blueprints": [ { "dependencies": [ { "arch": "noarch", "epoch": "0", "name": "2ping", "release": "2.el7", "version": "3.2.1" }, { "arch": "x86_64", "epoch": "0", "name": "acl", "release": "12.el7", "version": "2.2.51" }, { "arch": "x86_64", "epoch": "0", "name": "audit-libs", "release": "3.el7", "version": "2.7.6" }, { "arch": "x86_64", "epoch": "0", "name": "avahi-libs", "release": "17.el7", "version": "0.6.31" }, ... ], "modules": [ { "arch": "noarch", "epoch": "0", "name": "2ping", "release": "2.el7", "version": "3.2.1" }, { "arch": "x86_64", "epoch": "0", "name": "glusterfs", "release": "18.4.el7", "version": "3.8.4" }, ... ], "blueprint": { "description": "An example GlusterFS server with samba", "modules": [ { "name": "glusterfs", "version": "3.7.*" }, ... } } ] }
/api/v0/projects/list[?offset=0&limit=20]¶
List all of the available projects. By default this returns the first 20 items, but this can be changed by setting the offset and limit arguments.
Example:
{ "limit": 20, "offset": 0, "projects": [ { "description": "0 A.D. (pronounced "zero ey-dee") is a ...", "homepage": "http://play0ad.com", "name": "0ad", "summary": "Cross-Platform RTS Game of Ancient Warfare", "upstream_vcs": "UPSTREAM_VCS" }, ... ], "total": 21770 }
/api/v0/projects/info/<project_names>¶
Return information about the comma-separated list of projects. It includes the description of the package along with the list of available builds.
Example:
{ "projects": [ { "builds": [ { "arch": "x86_64", "build_config_ref": "BUILD_CONFIG_REF", "build_env_ref": "BUILD_ENV_REF", "build_time": "2017-03-01T08:39:23", "changelog": "- restore incremental backups correctly, files ...", "epoch": "2", "metadata": {}, "release": "32.el7", "source": { "license": "GPLv3+", "metadata": {}, "source_ref": "SOURCE_REF", "version": "1.26" } } ], "description": "The GNU tar program saves many ...", "homepage": "http://www.gnu.org/software/tar/", "name": "tar", "summary": "A GNU file archiving program", "upstream_vcs": "UPSTREAM_VCS" } ] }
/api/v0/projects/depsolve/<project_names>¶
Depsolve the comma-separated list of projects and return the list of NEVRAs needed to satisfy the request.
Example:
{ "projects": [ { "arch": "noarch", "epoch": "0", "name": "basesystem", "release": "7.el7", "version": "10.0" }, { "arch": "x86_64", "epoch": "0", "name": "bash", "release": "28.el7", "version": "4.2.46" }, { "arch": "x86_64", "epoch": "0", "name": "filesystem", "release": "21.el7", "version": "3.2" }, ... ] }
/api/v0/projects/source/list¶
Return the list of repositories used for depsolving and installing packages.
Example:
{ "sources": [ "fedora", "fedora-cisco-openh264", "fedora-updates-testing", "fedora-updates" ] }
/api/v0/projects/source/info/<source-names>¶
Return information about the comma-separated list of source names. Or all of the sources if '*' is passed. Note that general globbing is not supported, only '*'.
immutable system sources will have the "system" field set to true. User added sources will have it set to false. System sources cannot be changed or deleted.
Example:
{ "errors": [], "sources": { "fedora": { "check_gpg": true, "check_ssl": true, "gpgkey_urls": [ "file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora-28-x86_64" ], "name": "fedora", "proxy": "http://proxy.brianlane.com:8123", "system": true, "type": "yum-metalink", "url": "https://mirrors.fedoraproject.org/metalink?repo=fedora-28&arch=x86_64" } } }
POST /api/v0/projects/source/new¶
Add (or change) a source for use when depsolving blueprints and composing images.
The
proxyandgpgkey_urlsentries are optional. All of the others are required. The supported types for the urls are:
yum-baseurlis a URL to a yum repository.
yum-mirrorlistis a URL for a mirrorlist.
yum-metalinkis a URL for a metalink.If
check_sslis true the https certificates must be valid. If they are self-signed you can either set this to false, or add your Certificate Authority to the host system.If
check_gpgis true the GPG key must either be installed on the host system, orgpgkey_urlsshould point to it.You can edit an existing source (other than system sources), by doing a POST of the new version of the source. It will overwrite the previous one.
Example:
{ "name": "custom-source-1", "url": "https://url/path/to/repository/", "type": "yum-baseurl", "check_ssl": true, "check_gpg": true, "gpgkey_urls": [ "https://url/path/to/gpg-key" ] }
DELETE /api/v0/projects/source/delete/<source-name>¶
Delete a user added source. This will fail if a system source is passed to it.
The response will be a status response with status set to true, or an error response with it set to false and an error message included.
/api/v0/modules/list[?offset=0&limit=20]¶
Return a list of all of the available modules. This includes the name and the group_type, which is always "rpm" for lorax-composer. By default this returns the first 20 items. This can be changed by setting the offset and limit arguments.
Example:
{ "limit": 20, "modules": [ { "group_type": "rpm", "name": "0ad" }, { "group_type": "rpm", "name": "0ad-data" }, { "group_type": "rpm", "name": "0install" }, { "group_type": "rpm", "name": "2048-cli" }, ... ] "total": 21770 }
/api/v0/modules/list/<module_names>[?offset=0&limit=20]¶
Return the list of comma-separated modules. Output is the same as /modules/list
Example:
{ "limit": 20, "modules": [ { "group_type": "rpm", "name": "tar" } ], "offset": 0, "total": 1 }
/api/v0/modules/info/<module_names>¶
Return the module's dependencies, and the information about the module.
Example:
{ "modules": [ { "dependencies": [ { "arch": "noarch", "epoch": "0", "name": "basesystem", "release": "7.el7", "version": "10.0" }, { "arch": "x86_64", "epoch": "0", "name": "bash", "release": "28.el7", "version": "4.2.46" }, ... ], "description": "The GNU tar program saves ...", "homepage": "http://www.gnu.org/software/tar/", "name": "tar", "summary": "A GNU file archiving program", "upstream_vcs": "UPSTREAM_VCS" } ] }
POST /api/v0/compose¶
Start a compose. The content type should be 'application/json' and the body of the POST should look like this
Example:
{ "blueprint_name": "http-server", "compose_type": "tar", "branch": "master" }Pass it the name of the blueprint, the type of output (from '/api/v0/compose/types'), and the blueprint branch to use. 'branch' is optional and will default to master. It will create a new build and add it to the queue. It returns the build uuid and a status if it succeeds
Example:
{ "build_id": "e6fa6db4-9c81-4b70-870f-a697ca405cdf", "status": true }
/api/v0/compose/types¶
Returns the list of supported output types that are valid for use with 'POST /api/v0/compose'
Example:
{ "types": [ { "enabled": true, "name": "tar" } ] }
/api/v0/compose/queue¶
Return the status of the build queue. It includes information about the builds waiting, and the build that is running.
Example:
{ "new": [ { "id": "45502a6d-06e8-48a5-a215-2b4174b3614b", "blueprint": "glusterfs", "queue_status": "WAITING", "job_created": 1517362647.4570868, "version": "0.0.6" }, { "id": "6d292bd0-bec7-4825-8d7d-41ef9c3e4b73", "blueprint": "kubernetes", "queue_status": "WAITING", "job_created": 1517362659.0034983, "version": "0.0.1" } ], "run": [ { "id": "745712b2-96db-44c0-8014-fe925c35e795", "blueprint": "glusterfs", "queue_status": "RUNNING", "job_created": 1517362633.7965999, "job_started": 1517362633.8001345, "version": "0.0.6" } ] }
/api/v0/compose/finished¶
Return the details on all of the finished composes on the system.
Example:
{ "finished": [ { "id": "70b84195-9817-4b8a-af92-45e380f39894", "blueprint": "glusterfs", "queue_status": "FINISHED", "job_created": 1517351003.8210032, "job_started": 1517351003.8230415, "job_finished": 1517359234.1003145, "version": "0.0.6" }, { "id": "e695affd-397f-4af9-9022-add2636e7459", "blueprint": "glusterfs", "queue_status": "FINISHED", "job_created": 1517362289.7193348, "job_started": 1517362289.9751132, "job_finished": 1517363500.1234567, "version": "0.0.6" } ] }
/api/v0/compose/failed¶
Return the details on all of the failed composes on the system.
Example:
{ "failed": [ { "id": "8c8435ef-d6bd-4c68-9bf1-a2ef832e6b1a", "blueprint": "http-server", "queue_status": "FAILED", "job_created": 1517523249.9301329, "job_started": 1517523249.9314211, "job_finished": 1517523255.5623411, "version": "0.0.2" } ] }
/api/v0/compose/status/<uuids>[?blueprint=<blueprint_name>&status=<compose_status>&type=<compose_type>]¶
Return the details for each of the comma-separated list of uuids. A uuid of '*' will return details for all composes.
Example:
{ "uuids": [ { "id": "8c8435ef-d6bd-4c68-9bf1-a2ef832e6b1a", "blueprint": "http-server", "queue_status": "FINISHED", "job_created": 1517523644.2384307, "job_started": 1517523644.2551234, "job_finished": 1517523689.9864314, "version": "0.0.2" }, { "id": "45502a6d-06e8-48a5-a215-2b4174b3614b", "blueprint": "glusterfs", "queue_status": "FINISHED", "job_created": 1517363442.188399, "job_started": 1517363442.325324, "job_finished": 1517363451.653621, "version": "0.0.6" } ] }
DELETE /api/v0/compose/cancel/<uuid>¶
Cancel the build, if it is not finished, and delete the results. It will return a status of True if it is successful.
Example:
{ "status": true, "uuid": "03397f8d-acff-4cdb-bd31-f629b7a948f5" }
DELETE /api/v0/compose/delete/<uuids>¶
Delete the list of comma-separated uuids from the compose results.
Example:
{ "errors": [], "uuids": [ { "status": true, "uuid": "ae1bf7e3-7f16-4c9f-b36e-3726a1093fd0" } ] }
/api/v0/compose/info/<uuid>¶
Get detailed information about the compose. The returned JSON string will contain the following information:
id - The uuid of the comoposition
config - containing the configuration settings used to run Anaconda
blueprint - The depsolved blueprint used to generate the kickstart
commit - The (local) git commit hash for the blueprint used
deps - The NEVRA of all of the dependencies used in the composition
compose_type - The type of output generated (tar, iso, etc.)
queue_status - The final status of the composition (FINISHED or FAILED)
Example:
{ "commit": "7078e521a54b12eae31c3fd028680da7a0815a4d", "compose_type": "tar", "config": { "anaconda_args": "", "armplatform": "", "compress_args": [], "compression": "xz", "image_name": "root.tar.xz", ... }, "deps": { "packages": [ { "arch": "x86_64", "epoch": "0", "name": "acl", "release": "14.el7", "version": "2.2.51" } ] }, "id": "c30b7d80-523b-4a23-ad52-61b799739ce8", "queue_status": "FINISHED", "blueprint": { "description": "An example kubernetes master", ... } }
/api/v0/compose/metadata/<uuid>¶
Returns a .tar of the metadata used for the build. This includes all the information needed to reproduce the build, including the final kickstart populated with repository and package NEVRA.
The mime type is set to 'application/x-tar' and the filename is set to UUID-metadata.tar
The .tar is uncompressed, but is not large.
/api/v0/compose/results/<uuid>¶
Returns a .tar of the metadata, logs, and output image of the build. This includes all the information needed to reproduce the build, including the final kickstart populated with repository and package NEVRA. The output image is already in compressed form so the returned tar is not compressed.
The mime type is set to 'application/x-tar' and the filename is set to UUID.tar
/api/v0/compose/logs/<uuid>¶
Returns a .tar of the anaconda build logs. The tar is not compressed, but is not large.
The mime type is set to 'application/x-tar' and the filename is set to UUID-logs.tar
/api/v0/compose/image/<uuid>¶
Returns the output image from the build. The filename is set to the filename from the build with the UUID as a prefix. eg. UUID-root.tar.xz or UUID-boot.iso.
/api/v0/compose/log/<uuid>[?size=kbytes]¶
Returns the end of the anaconda.log. The size parameter is optional and defaults to 1Mbytes if it is not included. The returned data is raw text from the end of the logfile, starting on a line boundry.
Example:
12:59:24,222 INFO anaconda: Running Thread: AnaConfigurationThread (140629395244800) 12:59:24,223 INFO anaconda: Configuring installed system 12:59:24,912 INFO anaconda: Configuring installed system 12:59:24,912 INFO anaconda: Creating users 12:59:24,913 INFO anaconda: Clearing libuser.conf at /tmp/libuser.Dyy8Gj 12:59:25,154 INFO anaconda: Creating users 12:59:25,155 INFO anaconda: Configuring addons 12:59:25,155 INFO anaconda: Configuring addons 12:59:25,155 INFO anaconda: Generating initramfs 12:59:49,467 INFO anaconda: Generating initramfs 12:59:49,467 INFO anaconda: Running post-installation scripts 12:59:49,467 INFO anaconda: Running kickstart %%post script(s) 12:59:50,782 INFO anaconda: All kickstart %%post script(s) have been run 12:59:50,782 INFO anaconda: Running post-installation scripts 12:59:50,784 INFO anaconda: Thread Done: AnaConfigurationThread (140629395244800)
-
pylorax.api.v0.blueprint_exists(api, branch, blueprint_name)[source]¶ Return True if the blueprint exists
pylorax.api.workspace module¶
-
pylorax.api.workspace.workspace_delete(repo, branch, recipe_name)[source]¶ Delete the recipe from the workspace
-
pylorax.api.workspace.workspace_dir(repo, branch)[source]¶ Create the workspace's path from a Repository and branch