有性能问题，上HeapDump性能社区

JVM GC Log频繁刷STW日志：“Total time for which application threads were stopped”，打开SafePoint统计看到vmop都是ForceAsyncSafepoint。统计1s内STW日志，STW日志时间总计0.8580982s。

这是否表示1s有858ms都在STW吗？ForceAsyncSafepoint是什么含义呢？求大神帮忙解答~

使用ParNew+CMS，年轻代20G，老年代40G。查gc log发现程序特点是很少有晋升老年代的对象，老年代使用峰值不到200MB，所以堆大小配置有些不合理。

Log截取：

2020-11-23T15:23:05.028+0800: 3292.188: Total time for which application threads were stopped: 0.0281469 seconds, Stopping threads took: 0.0008531 seconds

2020-11-23T15:23:05.057+0800: 3292.217: Total time for which application threads were stopped: 0.0278397 seconds, Stopping threads took: 0.0006748 seconds

2020-11-23T15:23:05.091+0800: 3292.251: Total time for which application threads were stopped: 0.0282124 seconds, Stopping threads took: 0.0010456 seconds

2020-11-23T15:23:05.126+0800: 3292.286: Total time for which application threads were stopped: 0.0284797 seconds, Stopping threads took: 0.0011426 seconds

复制