...
- In order to get started, COmanage must first be registered as an application with GitHub. In order to manage group/team memberships, this must be done using the account of an owner of the GitHub organization containing the teams to be managed.
- Start by creating a new Provisioning Target of type GithubProvisioner.
- After it is created, the configuration page will appear. This page has the callback URL that you must use when registering with GitHub.
- After registering, populate the GitHub Username, GitHub Client ID, and GitHub Client Secret.
- Select the operations for the provisioner to handle.
- If removal of unknown members from GitHub teams is enabled, COmanage will keep the team membership exactly in sync with the corresponding COmanage group. This means Team members added directly via GitHub will (eventually) be removed if they do not correspond to a CO Group member.
- However, if removal of unknown members from GitHub teams is not enabled, (1) re-provisioning a group or (2) setting a group to Suspended will not remove any members from the corresponding GitHub team -- even for members who were formerly part of the CO Group. (Under normal operations, however, CO People will be removed from the GitHub team when a group membership is removed.)
- Upon clicking Save, GitHub may ask to authenticate and/or authorize COmanage.
- After completing authentication and authorization, COmanage will ask which GitHub organization to manage.
- COmanage will manage the memberships of GitHub Teams in the selected organization when those Teams have names that exactly match the names of COmanage groups within the CO.
In order to be provisioned to a GitHub Team, a CO Person must have an identifier of type GitHub (case sensitive) attached to their record, exactly matching the name of the CO Person's GitHub login.- If a Team member is not already a member of any Team within the GitHub Organization, GitHub will first send an invitation to the person before they become an active member of the team. (Soon, anyway. CO-942)
| Note | ||
|---|---|---|
| ||
In order to reduce the number of queries that count towards GitHub's API rate limit, the GitHub Provisioner uses a file based caching mechanism, with cache files written to This means that if multiple web servers are running COmanage Registry, each will have its own cache, effectively doubling the number of uncached queries. Under typical usage, this should not be a problem, though for heavier usage something like a database cache (CO-945) may be required. |
Connecting GitHub Usernames to CO People
...