ci: fix Windows E2E cluster setup being skipped

- Remove runner.os != 'Windows' guard from Setup E2E Tests — the step
  uses shell: bash (Git Bash on Windows) and already handles EXEC_CG
  for Windows; skipping it left no cluster for the tests to connect to
- Add || true to docker info | grep -i cgroup in setup-dual-e2e.sh so
  set -eo pipefail does not abort setup when cgroup info is absent on
  Windows Docker Desktop

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: JeffreyDallas

.github/workflows/zxc-e2e-test.yaml

@@ -393,7 +393,7 @@ jobs:
             cd ../solo
 
       - name: Setup E2E Tests
-        if: ${{ inputs.custom-job-label != 'One Shot Single - using Podman' && runner.os != 'Windows' && !cancelled() && !failure() }}
+        if: ${{ inputs.custom-job-label != 'One Shot Single - using Podman' && !cancelled() && !failure() }}
         uses: step-security/retry@e1d59ce1f574b32f0915e3a8df055cfe9f99be5d # v3.0.4
         with:
           max_attempts: 3

test/e2e/dual-cluster/setup-dual-e2e.sh

@@ -51,7 +51,7 @@ done
 
 docker network rm -f kind || true
 docker network create kind --scope local --subnet 172.19.0.0/16 --driver bridge
-docker info | grep -i cgroup
+docker info | grep -i cgroup || true
 
 # Setup Helm Repos
 helm repo add metrics-server https://kubernetes-sigs.github.io/metrics-server/ --force-update