多次大的gc，随后就会断连，这是因为什么呢？

日志如下
[2019-01-17T23:08:19,097][INFO ][o.e.m.j.JvmGcMonitorService] [node-142] [gc][101523] overhead, spent [322ms] collecting in the last [1s]
[2019-01-17T23:08:26,107][INFO ][o.e.m.j.JvmGcMonitorService] [node-142] [gc][101530] overhead, spent [370ms] collecting in the last [1s]
[2019-01-17T23:08:34,109][INFO ][o.e.m.j.JvmGcMonitorService] [node-142] [gc][101538] overhead, spent [259ms] collecting in the last [1s]
[2019-01-17T23:10:15,126][INFO ][o.e.d.z.ZenDiscovery     ] [node-142] master_left [{node-141}{3kWs-rFaR_yZVZOpRp1BFA}{nFj2LpDWS6qNIeEUKzpBLA}{192.168.30.141}{192.168.30.141:2810}], reason [failed to ping, tried [3] times, each with  maximum [1m] timeout]
[2019-01-17T23:10:15,127][WARN ][o.e.d.z.ZenDiscovery     ] [node-142] master left (reason = failed to ping, tried [3] times, each with  maximum [1m] timeout), current nodes: nodes:
   {node-143}{gIvN8KNiQsmY2Lsy20Expg}{TUXpFaUnRTKAa5Asdf19gA}{192.168.30.143}{192.168.30.143:2810}
   {node-141}{3kWs-rFaR_yZVZOpRp1BFA}{nFj2LpDWS6qNIeEUKzpBLA}{192.168.30.141}{192.168.30.141:2810}, master
   {node-142}{aYMFcfn3RvKDVfgncxC8cQ}{yodrtCrKSaWcVI49NE9dCg}{192.168.30.142}{192.168.30.142:2810}, local
 
几次断连都是这种表现（在断连之前会出现3-5次大的gc），从时间上讲，即使1s中有0.8s的gc，也不应该会导致断连，因为断连需要长达3min的时间无法访问，而gc总是间断性的；
给我的感觉，好像gc有误，把连接的那个对象当作garbage回收了一样。