Access Token for Dockerhub - authentication

I created a repository on hub.docker.com and now want to push my image to the Dockerhub using my credentials. I am wondering whether I have to use my username and password or whether I can create some kind of access token to push the docker image.
What I want to do is using the docker-image resource from Concourse to push an image to Dockerhub. Therefore I have to configure credentials like:
type: docker-image
source:
email: {{docker-hub-email}}
username: {{docker-hub-username}}
password: {{docker-hub-password}}
repository: {{docker-hub-image-dummy-resource}}
and I don't want to use my Dockerhub password for that.

In short, you can't. There are some solutions that may appeal to you, but it may ease your mind first to know there's a structural reason for this:
Resources are configured via their source and params, which are defined at the pipeline level (in your yml file). Any authentication information has to be defined there, because there's no way to get information from an earlier step in your build into the get step (it has no inputs).
Since bearer tokens usually time out after "not that long" (i.e. hours or days) which is also true of DockerHub tokens, the concourse instance needs to be able to fetch a new token from the authentication service every time the build runs if necessary. This requires some form of persistent auth to be stored in the concourse server anyway, and currently Dockerhub does not support CI access tokens a la github.
All that is to say, you will need to provide a username and password to Concourse one way or another.
If you're worried about security, there are some steps you can most likely take to reduce risk:
you can use --load-vars-from to protect your credentials from being saved in your pipeline, storing them elsewhere (LastPass, local file, etc).
you might be able to create a user on Dockerhub that only has access to the particular repo(s) you want to push, a "CI bot user" if you will.

Docker Hub supports Access Token
goto Account Settings > Security
its same as Github personal access token (PAT)
You can use this token instead of actual password

Related

GitHub API individual user mirrored permissions to only organization resources

TLDR -
I want users of my employer's organization who install my CLI tool to be able to use it to run commands that use the GitHub rest api that require permissions the user doesn't personally have, while restricting the user from using the CLI tool to perform unintended actions with the elevated permissions needed by the CLI tool. Is there a way to create a CLI tool that 1. uses permissions the users of the CLI tool don't have while at the same time 2. prevents those users from using the CLI tool's elevated permissions to perform unintended actions?
Extended -
I'm building a CLI tool for my employer's organization so that the organization members can automatically create a new branch and simultaneously add branch protection and create a PR. As far as I know, updating branch protection with the GitHub API requires admin access. However, not all of the users who use this CLI tool will personally have those permissions for their user account. But a user in the organization should still be able to use the CLI tool (which uses permissions the user potentially doesn't have).
I thought about using a personal access token so that the users aren't using a single set of centralized credentials (github app creds), but then the CLI tool would have access to at least one of the user's personal repositories and potentially wouldn't have the required permissions required by the CLI tool.
So is there a way to build a tool as a package that can be installed on a users machine that...
has permissions that the user of the tool potentially doesn't have and
prevents the user from doing anything unintended. e.g. the tool (which has admin permissions) can add branch protection rules, but the user shouldn't be able to use the tool's elevated permissions for anything else - e.g. getting the interaction limits for an organization
prevents the user from using the tool in any organization repo they don't have access to.
Ok, that's the gist of it, but I will go into more detail in case that helps.
So currently I'm using a GitHub app and giving it the required permissions (e.g. admin read/write for branch protection rules). I'm using octokit to authenticate to the GitHub API with the GitHub App. To do this, I'm passing in the app id, private key, and the installation id.
package structure
- .env
- authenticate_and_do_stuff.ts
authenticate_and_do_stuff.ts
// authenticate with GitHub app creds
// see https://github.com/octokit/octokit.js#authentication
const octokit = new Octokit({
authStrategy: createAppAuth,
auth: {
appId: process.env.GITHUB_APP_APP_ID,
privateKey: process.env.GITHUB_APP_PRIVATE_KEY,
installationId: process.env.GITHUB_APP_INSTALLATION_ID,
},
});
// create branch, add branch protection, etc.
octokit.rest.do stuff ...
This isn't ideal because then the user who installs the package needs those environment variables to be authenticated. And once they have those creds, they would be able to use them to do other things they potentially don't have permissions for (e.g. with the admin read/write permission the GitHub app has permissions for). So my thought is to have the user provide the authentication credentials that aren't the Github app credentials. This way, each user is using different credentials they provide. As for the reasons stated above, personal access tokens cannot be used. So the ideal situation is as follows -
A user installs the package - npm i my-github-cli-tool.
- node_modules
|
--- my-github-cli-tool
|
--- authenticate_and_do_stuff.ts
|
- some_custom_app_file.ts
- .env.cli_tool
They generate credentials and put them into a custom env file e.g. .env.cli_tool.
They use the CLI tool to create a new branch in an organization repo they have access to, create the branch protection rules (with the admin read/write access that the user doesn't have, but the credentials in step 2 has).
If the user tries to alter the cli tool or use the credentials in step 2 to use it to access repos they don't have access to or perform other actions with the admin read/write permissions, they shouldn't be able to.
If I stick with using the GitHub app, the source code on the user's machine has to get the credentials somehow. And those credentials need to have elevated permissions that the user might not have. So if these credentials are exposed to the source code on the user's machine, doesn't that mean the user can just grab those credentials and use them for nefarious purposes?

GitHub: SSH over HTTPS for third party app

I am a hardware guy first and software second so GitHub is not my forte.
I had Altium Designer setup with my GitHub server for version control. When GitHub forced 2FA recently it broke the link to Altium which, unfortunately doesn't have stellar GitHub integration.
There are 6 fields I am allowed to enter in Altium to point it to my (GitHub) server:
1.) Method (HTTP, HTTPS, file, svn)
2.) Server (URL)
3.) Port
4.) Repo Subfolder
5.) username
6.) Password
Again, nothing changed except moving to 2FA. Now, when I attempt to login it obviously says it could not connect to the server because Altium has no provisions to provide a token during the login process.
I read the article at GitHub here: https://docs.github.com/en/free-pro-team#latest/github/authenticating-to-github/using-ssh-over-the-https-port
But I have no idea if that will do anything for me. Is there a way to route my Altium server connection to use my SSH key outside of the Altium environment? Or perhaps another way to "whitelist" my desktop in GitHub for SSO?
GitHub has not forced 2FA on for users. That wouldn't be useful, because people could just not set up a second factor. It's possible your organization has required this, though.
However, GitHub is deprecating the use of a plain password when using Git over HTTPS in favor of a token. Using a plain password was already forbidden for users who use 2FA, since there's no place to send a 2FA code (and for automated systems, doing that would be very inconvenient).
In this case, it's easy to keep using HTTPS: just generate a personal access token (in the developer settings) with the repo scope and paste it into the password field. Git doesn't know the difference between a password and a token; they're both the same to it. This also has a bunch of other benefits:
If you change your password, the token isn't automatically cleaned up, so you don't have to change Altium Designer.
If you decide you want to revoke that token, you can do so independently of changing your password.
If you're using SSO, you need to enable that token for SSO using the drop-down before it can be used to access protected resources.

How to use Github Personal Access Token in Jenkins

I can ask this question in many ways, like
How to configure Jenkins credentials with Github Personal Access Token
How to clone Github repo in Jenkins using Github Personal Access Token
So this is the problem
The alternate solution that I am aware of
SSH connection
username password configuration in Jenkins. However,
use of a password with the GitHub API is now deprecated.
But My question is how to setup Github connection with Jenkins using Personal Access Token
[UPDATE]
The new solution proposed by git is
https://github.blog/2020-12-15-token-authentication-requirements-for-git-operations/
Which says:
Beginning August 13, 2021, we will no longer accept account passwords
when authenticating Git operations and will require the use of
token-based authentication, such as a personal access token (for
developers) or an OAuth or GitHub App installation token (for
integrators) for all authenticated Git operations on GitHub.com. You
may also continue using SSH keys where you prefer.
What you need to do:
https://github.blog/2020-12-15-token-authentication-requirements-for-git-operations/#what-you-need-to-do-today
Basically, change the add URL as
https://<access token>#github.com/<userName>/<repository>.git
Something like this
https://<access token>#github.com/dupinder/NgnixDockerizedDevEnv.git
and set the credentials to none.
Thanks to #Gil Stal
[OLD Technique]
After many discussion on multiple threads from Stackoverflow
I found one thread that is useful.
Refer to this answer:
https://stackoverflow.com/a/61104603/5108695
Basically
Personal access token can be used as a password, as far as Jenkins is concerned at least. I added new credentials to the credential manager.
Go to Jenkins
Go to credentials > System > Global credentials > Add credentials a page will open.
In Kind drop-down select Username and password.
In User put a non-existing username like jenkins-user or user.
Add Personal Access Token in the password field
Now start configuring your project.
source code management tab, select new configured credentials from Drop-down near credential Under Repository URL
So this is how we can configure or setup Authentication between Jenkins and Github using Personal Access Token
References:
Git Clone in Jenkins with Personal Access Token idles forever
Change jenkins pipeline to use github instead of gitlab
The accepted answer wont work anymore because of this: https://github.blog/2020-12-15-token-authentication-requirements-for-git-operations.
You will need to:
Change the URL of the repo to: https://<access token>#github.com/<user-name>/<repo-name>.git (Replace every <...> with the real parameters)
Set the credentials to none.
As of August 2021 the answer posted by Dupinder Singh is accurate. The only thing I would add is that if you are part of a team, the url format appears to be a bit different. This is what worked for me:
https://<access token>#github.com/<team>/<repo>.git
for example
https://ghp_6dh3jdk394jsmbh299jjdg20fh87hd83ksk39#github.com/MyKuleTeam/KuleGuyCode.git
Note that if you use a personal access token you don't need to have any github credentials stored in jenkins.
As for credentials for Jenkins Github Plugin, please be aware only Personal access tokens are now accepted by this plugin.
To generate such a token, follow the Github docs (e.g. here). Don't save it, it can be regenerated in Github and updated in Jenkins if lost or when migrating to a different server.
To add the token do Jenkins credentials store, go to <JENKINS_URL:PORT>/credentials/store/system/domain/_/newCredentials and select Kind "Secret text" (not the default "Username and password"), then paste the token as Secret and choose some ID.
Testing: the credential should appear on the list of Credentials at <JENKINS_URL:PORT>/credentials/ and be selectable from the drop-down list at <JENKINS_URL:PORT>/configure/, where pressing the "Test connection" button should display "Credentials verified for user <GITHUB_USER>".
More info: see the Github plugin docs.
Caveats: Git Plugin has its long-standing issues, so if the newly created "Secret text" does not appear in your pipelines, try if this solution helps (with "the user who triggered the build" considered safer than "SYSTEM"):
client-and-managed-masters/why-credentials-are-not-listed-in-the-git-scm-section
There is (yet another) way to do this as of 2020/04 which is supposed to be superior to personal access tokens. The best part is that you can continue using a username/password-style credential, and the plugin will handle authenticating with GitHub in the background.
Benefits include:
Larger rate limits - The rate limit for a GitHub app scales with your organization size, whereas a user based token has a limit of 5000 regardless of how many repositories you have.
User-independent authentication - Each GitHub app has its own user-independent authentication. No more need for 'bot' users or figuring out who should be the owner of 2FA or OAuth tokens.
Improved security and tighter permissions - GitHub Apps offer much finer-grained permissions compared to a service user and its personal access tokens. This lets the Jenkins GitHub app require a much smaller set of privileges to run properly.
Access to GitHub Checks API - GitHub Apps can access the the GitHub Checks API to create check runs and check suites from Jenkins jobs and provide detailed feedback on commits as well as code annotation
Links:
https://www.jenkins.io/blog/2020/04/16/github-app-authentication/
https://github.com/jenkinsci/github-branch-source-plugin/blob/master/docs/github-app.adoc

CloudFoundry CLI login not working (Credentials were rejected, please try again)

So far I have always been able to log in successfully via sso.
cf login -a url --sso
I need another way to log in for my pipeline script and tried the following command.
cf login [-a API_URL] [-u USERNAME] [-p PASSWORD] [-o ORG] [-s SPACE]
This command does not work with my user, nor with a technical user to whom all necessary roles have been assigned (M D A). I get the following message.
API endpoint: url
Password>
Authenticating...
Credentials were rejected, please try again.
Does anyone know how to solve this problem?
Or maybe an alternative to create a gradle task, for example, that can be executed in a jenkins pipeline.
At the end, I want to automate a deploy (to cloud) of an artifact with my Jenkins pipeline.
You provided —sso flag, so you shouldn’t see a password prompt. Instead you should be given the url to get a token.
Maybe your CF has been misconfigured and does not support SSO yet. I tried to fix the CF CLI to avoid this but it was oddly rejected https://github.com/cloudfoundry/cli/pull/1624
Try fixing your CF installation (it needs to provide some prompts), or skip the —sso flag usage.
Using --sso and -u/-p are not doing the same thing on the backend, and there's no guarantee that a user which can login through SSO is also set up to login as a user stored directly in UAA. UAA has multiple origin's from which users can be loaded, like SAML, LDAP and internal to UAA. When you use the --sso flag, you are typically logging in via a user from your company's SAML provider. When you use the -u/-p flags, it's typically LDAP or UAA, something UAA validates directly.
In order for what you are trying to do to work, you would need to have a user available with an origin in SAML (for --sso) and a user in origin LDAP or UAA (internal), and technically those would be two separate users (despite the fact that they may have the same credentials).
At any rate, if you normally login with the --sso flag and you want to automate work, what you really want is to get a UAA client that is set with the grant type of client credentials. You can then use cf auth CLIENT_ID CLIENT_SECRET --client-credentials to automate logging in.
Typically you don't want your user account to be tied to pipelines and automated scripts anyway. If you leave the company and your user get deactivated then everything breaks :) You want a service account, and that is basically a client enabled with the client credentials grant type in UAA.

How to use LDAP credentials offline?

I would like to use an LDAP server (probably Apache directory) to manage logins and credentials for an application. From time to time the application needs to work offline (on a laptop) without a connection to the LDAP server.
What is the best way to replicate the credentials localy?
I have already thought about:
Using Mitosis to replicate the LDAP server on the laptop.
But it would be a quite "heavy" and complicated solution. Moreover Mitosis seems not be be finished yet.
Exporting the credentials as LDIF file that could be stored on the laptop.
But I would need a way to check that the LDIF file actually comes from the LDAP server (The file should include a kind of signature). Moreover I would like to reject LDIF files that haven't be updated for more than a week. It would be nice if I could avoid implementing signing and age check myself.
Any other ideas or tools that could help me?
Edited Edit: I had a look at Kerberos because the documentation of the Java-Kerberos-API seems to say that it is possible to use a cached ticket in a local cache and I thought this might be a solution for me. Moreover Kerberos can be added as plugin to Apache Directory.
But the Kerberos cache stores decrypted tickets (aiming at sharing them with other applications). I would need the crypted version of the ticket to be able to check the user password during an offline session. Conclusion: Kerberos doesn't offer a simple solution to my problem.
Knowing that it will be probably ok if the user have to log on once online before being able to log on offline, consider the following algorithm:
user provides your application with a (username + password)
application attempts to contact LDAP for authentication
working online? (e.g. connection successful)
application authenticates against LDAP using (username + password)
authentication succesful?
application stores or updates hash(password) as (cached_credentials) for (username) into local secure storage
application proceeds as authenticated [[STOP]]
authentication failed?
application proceeds as non-authenticated (incorrect credentials) [[STOP]]
working offline? (e.g. network error)
application attempts retrieve (cached_credentials) for (username) from local secure storage
(cached_credentials) exists AND more recent than (1 week)?
application compares (cached_credentials) against hash(password)
match?
application proceeds as authenticated [[STOP]]
no match?
application proceeds as non-authenticated (incorrect credentials) [[STOP]]
(cached_credentials) does not exist OR less recent than (1 week)?
application proceeds as non-authenticated (network error) [[STOP]]
This is (or was, IIRC), by the way, the same model employed by Windows NT+ for user authentication against domain controllers. Upon login an attempt is made to authenticate against the domain controller and create or update the local (cached) version of the user profile. If the domain controller is not available, the user is prompted to proceed with authentication against the credentials captured in the local (cached) profile (if one exists.)
EDIT
Yes, this is, in spirit, the same solution as copying an ldif file locally, except that you do not have to parse ldif when you're offline. :)
It is understood that you can store any additional attributes (permissions, etc.) in your cache
It is also understood that 'secure storage' is at least signed. :) You can do this easily enough with a SHA-1 hash and a secret, or you can use full-fledged cryptographic providers available on your platform (or in Java, if using Java.) You do not need to crypt it as long as no secret information is stored inside.
Here is the solution I decided to use (I have already described it in an edit to my question, but I would like to able to accept an answer to "close" the question):
As I have not found another solution, I decided to use an LDIF export, add a timestamp as comment at the beginning of the file and then sign the file. To sign the file I calculate an hash value (SHA-1) of the file + a secret key. The signature is added as comment at the beginning of the file. To check the signature I remove the first line of the signed file and recalculate the hash value.