現象：
docker swarm集群中節點的上的容器不斷的up down
查看日志，如下：
Feb 26 11:43:16 sz-drip dockerd[5065]: time="2020-02-26T11:43:16.124650220+08:00" level=error msg="heartbeat to manager { } failed" error="rpc error: code = NotFound desc = node not register
Feb 26 11:43:16 sz-drip dockerd[5065]: time="2020-02-26T11:43:16.124764380+08:00" level=error msg="agent: session failed" backoff=100ms error="node not registered" module=node/agent node.id=
Feb 26 11:43:16 sz-drip dockerd[5065]: time="2020-02-26T11:43:16.124803737+08:00" level=info msg="manager selected by agent for new session: { }" module=node/agent node.id=a5drrk7eo59x4m6itl
Feb 26 11:43:16 sz-drip dockerd[5065]: time="2020-02-26T11:43:16.124835689+08:00" level=info msg="waiting 47.652971ms before registering session" module=node/agent node.id=a5drrk7eo59x4m6itl
Feb 26 11:43:16 sz-drip dockerd[5065]: time="2020-02-26T11:43:16.278198549+08:00" level=info msg="worker a5drrk7eo59x4m6itlbsf28pq was successfully registered" method="(*Dispatcher).register
原因：
docker swarm 節點間通信默認超時時間為 5s，在虛擬化環境下或者網絡不太好的情況下有時候會超時，導致節點重啟。
方案：
調高swarm心跳檢測時間。
docker swarm update --dispatcher-heartbeat 60s #設置docker swarm的心跳檢測時間為1分鍾，在manages節點執行