After testing many different JVM GC settings and conducting a large number of tests where I had problems with long large GC pauses, I am now testing the G1GC JVM GC. In addition to this, I also collect data with a performance monitor, and only applications that work (besides system services, ...) are the GlassFish server with my application. I did not find anything strange in the performance monitoring protocol (CPU usage is about 5-10%, and it gets a little higher when GC happens, memory is about 60%, ...). This is the fifth day of testing, and I noticed the following:
, () GC , ( GC 20 , GC 160000 /,...). GC 2 ( - 150 , ), GC , (. , () GC) GC 12000 / . , GC, , / , . , . : , , GC 5000 / ( GC 50-100 ), - . , GC 3 , , , . ( 150 , 80 )? , , GC ( , ).
EDIT: - GC, GCViewer, GCViewer:
GC:
2015-06-08T08:09:13.123+0200: 572815.533: [GC concurrent-root-region-scan-start]
2015-06-08T08:09:13.139+0200: 572815.560: [GC concurrent-root-region-scan-end, 0.0271771 secs]
2015-06-08T08:09:13.139+0200: 572815.560: [GC concurrent-mark-start]
2015-06-08T08:09:16.302+0200: 572818.721: [GC concurrent-mark-end, 3.1612900 secs]
2015-06-08T08:09:16.318+0200: 572818.729: [GC remark 572818.729: [Finalize Marking, 0.0002590 secs] 572818.729: [GC ref-proc, 0.4479462 secs] 572819.177: [Unloading, 3.2004912 secs], 3.6499382 secs]
[Times: user=0.20 sys=0.08, real=3.64 secs]
, , user + sys, 3 .