{Heap before GC invocations=4906 (full 11):
par new generation total 4587520K, used 4585788K [0x00000002c0000000, 0x0000000400000000, 0x0000000400000000)
eden space 3932160K, 99% used [0x00000002c0000000, 0x00000003afe4f068, 0x00000003b0000000)
from space 655360K, 100% used [0x00000003d8000000, 0x0000000400000000, 0x0000000400000000)
to space 655360K, 0% used [0x00000003b0000000, 0x00000003b0000000, 0x00000003d8000000)
concurrent mark-sweep generation total 15728640K, used 13605986K [0x0000000400000000, 0x00000007c0000000, 0x00000007c0000000)
Metaspace used 161034K, capacity 168360K, committed 169020K, reserved 1202176K
class space used 15924K, capacity 17053K, committed 17276K, reserved 1048576K
2016-10-12T12:46:56.321+0800: 78450.842: [GC (Allocation Failure) 2016-10-12T12:46:56.322+0800: 78450.843: [ParNew: 4585788K->655360K(4587520K), 2.6300860 secs] 18191774K->15487368K(20316160K), 2.6311850 secs] [Times: user=34.75 sys=1.59, real=2.63 secs]
Heap after GC invocations=4907 (full 11):
par new generation total 4587520K, used 655360K [0x00000002c0000000, 0x0000000400000000, 0x0000000400000000)
eden space 3932160K, 0% used [0x00000002c0000000, 0x00000002c0000000, 0x00000003b0000000)
from space 655360K, 100% used [0x00000003b0000000, 0x00000003d8000000, 0x00000003d8000000)
to space 655360K, 0% used [0x00000003d8000000, 0x00000003d8000000, 0x0000000400000000)
concurrent mark-sweep generation total 15728640K, used 14832008K [0x0000000400000000, 0x00000007c0000000, 0x00000007c0000000)
Metaspace used 161034K, capacity 168360K, committed 169020K, reserved 1202176K
class space used 15924K, capacity 17053K, committed 17276K, reserved 1048576K
}
2016-10-12T12:47:00.136+0800: 78454.657: [CMS-concurrent-abortable-preclean: 13.471/16.947 secs] [Times: user=87.65 sys=8.63, real=16.95 secs]
gc时间超过了设置的60s zookeeper session timeout。导致了服务不可用（这里面就不详细描述了）