Fix Ansible playbook setup issues: OAuth redirect URIs, Docker version pinning, and CoreDNS external resolution
### Issue Title:
```text
Fix local MEC Sandbox Ansible setup issues: OAuth redirect URIs, Docker version pinning, and CoreDNS external resolution
```
### Issue Description:
The following setup issues were identified during local MEC Sandbox deployments using the Ansible playbooks:
1. **OAuth redirect URIs in `.meepctl-repocfg.yaml` not updated**
- **Cause**: The playbook task uses the regex `host:\s*mec-platform\.etsi\.org` and `redirect-uri:\s*https://mec-platform\.etsi\.org/...` to configure hostnames. However, the default config template in `.meepctl-repocfg.yaml` actually uses `try-mec.etsi.org` as the host. This mismatch prevents the Ansible host replacement from executing, leaving the callback URL as `try-mec.etsi.org` rather than updating to the sandbox's actual host IP/address, which causes GitHub/GitLab OAuth logins to fail.
- **Fix**: Update the playbook regex in `playbooks/roles/mec_sandbox/mec_config/tasks/main.yml` to look for `try-mec.etsi.org` or to support either host placeholder.
2. **Docker engine package version pinning causes deployment failures**
- **Cause**: The playbook pins Docker to `5:29.2.1-1~ubuntu...` in `playbooks/inventories/dev/group_vars/all.yml`. If the target host already has a newer version of Docker (e.g. 29.5.0) installed and held, `apt` refuses to install or downgrade it without `--allow-downgrades`.
- **Fix**: Remove the version pin from the `apt` task in `playbooks/roles/docker/tasks/install.yml` and clean up unused variables from `playbooks/inventories/dev/group_vars/all.yml` so it installs the latest or keeps the existing compatible version.
3. **CoreDNS external resolution failure inside Kubernetes pods**
- **Cause**: The host uses `systemd-resolved` (default on modern Ubuntu), meaning `/etc/resolv.conf` on the host contains `nameserver 127.0.0.53`. CoreDNS inside the Kubernetes cluster inherits this address, which points to a loopback address and fails from inside the pods.
- **Fix**: Patch the CoreDNS ConfigMap to forward requests to public DNS resolvers (8.8.8.8, 1.1.1.1) instead of using the host `/etc/resolv.conf`.
issue