The epel-testing repository contains updates scheduled to be released for the maintained releases of EPEL. User testing and feedback provided via Bodhi, on the epel-devel mailing list and the relevant Bugzilla is vital to ensure that good updates are released quickly and bad ones kept away from release.
Using the epel-testing repository
Enabling the repository permanently
The following command will enable the updates testing repository permanently
yum-config-manager --enable epel-testing
Use yum repolist to verify. If you wish to disable it again, run the following command
yum-config-manager --disable epel-testing
Tip: yum distro-sync will sync the packages to the versions available in the repository and might be useful to run after you disable the testing repository to downgrade packages back to the stable versions.
Note that yum-config-manager command is available as part of yum-utils package and should be installed by default, at least on RHEL >= 6. On older releases, or if yum-config-manager is not available, you may need to edit the
/etc/yum.repos.d/epel-testing.repo file manually and set
Enabling the repository temporarily
If you'd rather not enable the epel-testing repository permanently but just use it on a case-by-case basis, you can do this with yum. The command:
yum update --enablerepo=epel-testing
will update the entire system using packages from the epel-testing repository, while the command:
yum install <foo> --enablerepo=epel-testing
will install or update only the package named <foo> from the epel-testing repository.
What to test, testing, and reporting results
The Bodhi system is used to track and collate feedback on testing updates. All testing updates will be shown in the Bodhi system. First of all, if any test update package works worse for you in any respect than the pre-update version did, this is a problem that should be communicated to the developers. Secondly, when you click on a certain update, you will see a screen with more information on the update. The Details section should give you information on what the update is intended to fix. You should, if possible, test that the update does indeed fix the issues it claims to fix.
You can give your feedback on a test update by using the Bodhi web interface. There is a Login link in the left-hand sidebar. Log in using your Fedora account. If you don't have a Fedora account, you can create an account here. Once you are logged in, you will be able to leave a comment on the update. Underneath the comment box are three options: Untested, Works for me, and Does not work. For a guide on when to leave each type of feedback, read the update feedback guidelines.
Each Works for me adds 1 to the test update's karma, while each Does not work subtracts 1 from it. Untested leaves the karma unchanged. Usually, test updates with karma of 3 are automatically sent out as full official updates, while test updates with karma of -3 are automatically withdrawn from the testing repository. As you can see, your testing and feedback is vital to make sure that good updates are released quickly and bad ones don't get out to the general public.
Note on using DNF with epel-testing and dependency errors (EPEL 7 only)
If you use the DNF tool rather than yum, note that unlike yum, it does not notify you of dependency errors in the set of available updates by default. yum will, by default, print some information about the dependency errors, and then fail the update: you have to run
yum update --skip-broken to install all the updates that do not have dependency errors. DNF essentially does
--skip-broken by default, and silently - it will install all updates that do not have dependency errors, and not tell you about those that do.
To use dnf to spot updates with dependency errors, you can run
dnf update --best, which should act like yum's default behaviour.