CoreOSdocker拉片

我有以下单位：

[Unit] Description=Bootstrap Docker Swarm cluster Requires=docker.service coreos-metadata.service After=docker.service coreos-metadata.service [Service] Type=oneshot RemainAfterExit=true Environment="AWSCLI_IMAGE=quay.io/coreos/awscli:latest" EnvironmentFile=/run/metadata/coreos ExecStartPre=/usr/bin/systemctl is-active docker.service ExecStartPre=-/usr/bin/docker pull $AWSCLI_IMAGE ExecStartPre=-/usr/bin/docker run --rm --net=host \ -v /run/metadata:/run/metadata \ -w /run/metadata \ $AWSCLI_IMAGE aws s3api get-object \ --region ${bucket_region} \ --bucket ${bucket_name} --key leader_ipv4_local \ leader_ipv4_local ExecStart=/opt/bin/init-swarm.sh [Install] WantedBy=multi-user.target 

当它运行时，在第一次运行时，以一种随机意外的方式，我得到以下错误：

 Dec 26 13:14:49 ip-192-168-4-70 docker[958]: Error response from daemon: Get https://quay.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers) Dec 26 13:14:49 ip-192-168-4-70 docker[965]: Unable to find image 'quay.io/coreos/awscli:latest' locally Dec 26 13:15:04 ip-192-168-4-70 docker[965]: /run/torcx/bin/docker: Error response from daemon: Get https://quay.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers). Dec 26 13:15:04 ip-192-168-4-70 docker[965]: See '/run/torcx/bin/docker run --help'. 

当它第二次运行，然后拉成功。 任何想法？