date: '2021-03-16'
sections:
  security_fixes:
    - Packages have been updated to the latest security versions.
  bugs:
    - Systemd journal logs were duplicated in multiple places.
    - A site admin could get a 500 error page while trying to view issues referenced from private repositories.
    - Importing of repository archives from GitHub Enterprise Server that are missing repository files would fail with an error.
  known_issues:
    - On a freshly set up GitHub Enterprise Server without any users, an attacker could create the first admin user.
    - Custom firewall rules are not maintained during an upgrade.
    - Git LFS tracked files [uploaded through the web interface](https://github.com/blog/2105-upload-files-to-your-repositories) are incorrectly added directly to the repository.
    - Issues cannot be closed if they contain a permalink to a blob in the same repository where the file path is longer than 255 characters.
    - When "Users can search GitHub.com" is enabled with GitHub Connect, issues in private and internal repositories are not included in GitHub.com search results.
    - |
      Users may experience assets such as avatars not loading, or a failure to push/pull code. This may be caused by a PID mismatch in the `haproxy-cluster-proxy` service. To determine if you have an affected instance:

      **Single instance**

      1. Run this in the [administrative shell](https://docs.github.com/en/enterprise-server/admin/configuration/accessing-the-administrative-shell-ssh) (SSH):

        ```
        if [ $(cat /var/run/haproxy-cluster-proxy.pid) -ne $(systemctl show --property MainPID --value haproxy-cluster-proxy) ]; then echo 'Main PID of haproxy-cluster-proxy does not match /var/run/haproxy-cluster-proxy.pid'; fi
        ```

      2. If it shows that there is a mismatch, reboot the instance.

      **Cluster or High Availability configuration**

      1. Run this in the [administrative shell](https://docs.github.com/en/enterprise-server/admin/configuration/accessing-the-administrative-shell-ssh) (SSH):

        ```
        ghe-cluster-each -- 'if [ $(cat /var/run/haproxy-cluster-proxy.pid) -ne $(systemctl show --property MainPID --value haproxy-cluster-proxy) ]; then echo 'Main PID of haproxy-cluster-proxy does not match /var/run/haproxy-cluster-proxy.pid'; fi'
        ```

      2. If it shows one or more nodes are affected, reboot the affected nodes.
    - When a replica node is offline in a high availability configuration, {% data variables.product.product_name %} may still route {% data variables.product.prodname_pages %} requests to the offline node, reducing the availability of {% data variables.product.prodname_pages %} for users.