chore(ci): dont detach ebs (#6441)

Seeming some EBS corruption and correlates to multiple runners starting,
could be detaching someone who just attached

.github/ci-setup-action/action.yml

@@ -59,7 +59,7 @@ runs:
       if: ${{ inputs.concurrency_key }}
       with:
         run: |
-          while [ -f "${{ inputs.concurrency_key }}.lock" ]; do sleep 1 ; echo "Lock is currently held by $(cat '${{ inputs.concurrency_key }}.lock'), waiting..." ; done
+          while [ -f "/home/ubuntu/${{ inputs.concurrency_key }}.lock" ]; do sleep 1 ; echo "Lock is currently held by $(cat '/home/ubuntu/${{ inputs.concurrency_key }}.lock'), waiting..." ; done
           echo "${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}" > "/home/ubuntu/${{ inputs.concurrency_key }}.lock"
           echo "/home/ubuntu/${{ inputs.concurrency_key }}.lock acquired."
         post: |

scripts/ci/attach_ebs_cache.sh

@@ -111,11 +111,6 @@ while [ "$(aws ec2 describe-volumes \
   elapsed_time=$((elapsed_time + WAIT_INTERVAL))
 done
 
-# First, make sure this is detached from any instances stuck in stopping state
-aws ec2 detach-volume \
-  --region $REGION \
-  --volume-id $VOLUME_ID || true
-
 # Attach volume to the instance
 aws ec2 attach-volume \
   --region $REGION \
@@ -162,4 +157,4 @@ rm -rf /var/lib/docker/tmp-images
 systemctl restart docker
 # important: everything (except earthly ls) should go through earthly-ci
 scripts/earthly-ci bootstrap
-touch /home/ubuntu/.setup-complete
+touch /home/ubuntu/.setup-complete