RFR: 8059216: Make PrintGCApplicationStoppedTime print information about starting threads

Vladimir Kempik vladimir.kempik at oracle.com
Mon Oct 20 14:32:55 UTC 2014


Hi,

Please review patch for improvement in PrintGCApplicationStoppedTime 
logging:

http://cr.openjdk.java.net/~vkempik/8059216/webrev.00
https://bugs.openjdk.java.net/browse/JDK-8059216

Sometimes GC kicks in much later than safepoint start time.
And  without -XX:+PrintGCApplicationConcurrentTime it's may be not clear 
why that happens.
The patch adds more logging to PrintGCApplicationStoppedTime to show how 
much time it took to stop all threads.

Example of log after applying patch:

0,153: Application time: 0,0431575 seconds    (printed by 
-XX:+PrintGCApplicationConcurrentTime)
0,619: #1: [GC remark 0,619: #1: [Finalize Marking, 0,0001732 secs]
0,619: #1: [GC ref-proc, 0,0001122 secs] 0,619: #1: [Unloading, 
0,0006922 secs]
0,620: #1: [GC aggregate-data, 0,0004849 secs], 0,0015962 secs]
  [Times: user=0,00 sys=0,00 real=0,00 secs]
0,620: Total time for which application threads were stopped: 0,4673289 
seconds, Stopping threads took: 0,4656409 seconds

Testing: JPRT.

Thanks,
Vladimir.


More information about the hotspot-runtime-dev mailing list