docs: tell how to push one tag; that rc start with n=0 (#3222)

If your local repo tags don't match the remote, then `git push --tags`
will push _all_
tags. This confuses the release workflow and it doesn't trigger
properly. It can also
push junk tags and trigger an accidental release (hence how a 0.1
release showed up
months ago; I accidentally pushed a junk tag).

Along the way, mention that N=0 to start with for RCs

Author: rickeylev

RELEASING.md

@@ -28,18 +28,19 @@ The next step is to create tags to trigger release workflow, **however**
 we start by using release candidate tags (`X.Y.Z-rcN`) before tagging the
 final release (`X.Y.Z`).
 
-1. Create release candidate tag and push. Increment `N` for each rc.
+1. Create release candidate tag and push. The first RC uses `N=0`. Increment
+   `N` for each RC.
    ```
-   git tag X.Y.0-rcN upstream/release/X.Y && git push upstream --tags
+   git tag X.Y.0-rcN upstream/release/X.Y && git push upstream tag X.Y.0-rcN
    ```
 2. Announce the RC release: see [Announcing Releases]
 3. Wait a week for feedback.
    * Follow [Patch release with cherry picks] to pull bug fixes into the
      release branch.
    * Repeat the RC tagging step, incrementing `N`.
 4. Finally, tag the final release tag:
-   ```
-   git tag X.Y.0 upstream/release/X.Y && git push upstream --tags
+   ```shell
+   git tag X.Y.0 upstream/release/X.Y && git push upstream tag X.Y.0
    ```
 
 Release automation will create a GitHub release and BCR pull request.