JDK-6596629 : System hangs/freezes for a period during a long running test of Java Application - GC logs show gaps
  • Type: Bug
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 5.0u12
  • Priority: P2
  • Status: Closed
  • Resolution: Duplicate
  • OS: linux_redhat_4.0
  • CPU: x86
  • Submitted: 2007-08-23
  • Updated: 2010-04-02
  • Resolved: 2007-08-30
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
Other
5.0u12Resolved
Related Reports
Duplicate :  
Description
Oracle reports system freeze/hang during stress testing of their OCMS server for a period of 15-30 seconds. The tests run for a period of upto 120 hours.

System has enough memory and only a portion of it is allocated to heap. 

Details are provided as follows - the first set of JVM options are the default and the second set are the tuned options.

I have already suggested testing with JDK6 and comparing the results.

-------------------------------------------------------------------

JDK: 1.5.0_12


JVM settings 1: 

-Xloggc:$ORACLE_HOME/gc.log 
-XX:+PrintGCDetails 
-XX:NewRatio=3
-XX:+UseConcMarkSweepGC 
-XX:+UseParNewGC 
-XX:PermSize=128m
-XX:MaxPermSize=128m 
-Xss128k 
-Xmx1280M 
-Xms1280M


JVM settings 2: 

-Xloggc:$ORACLE_HOME/gc.log 
-XX:+PrintGCDetails
-XX:NewSize=32m 
-XX:MaxNewSize=32m 
-XX:+UseConcMarkSweepGC
-XX:+CMSIncrementalMode
-XX:+CMSIncrementalPacing 
-XX:CMSIncrementalDutyCycleMin=0
-XX:CMSIncrementalDutyCycle=10 
-XX:+UseTLAB 
-XX:+UseParNewGC
-XX:MaxTenuringThreshold=0 
-XX:SurvivorRatio=128
-XX:+UseCMSCompactAtFullCollection 
-XX:CMSFullGCsBeforeCompaction=0
-XX:PermSize=64m 
-XX:MaxPermSize=64m
-Xss128k
-Xmx1280M 
-Xms1280M 


Configiguration:


Server: (2 CPU, 3.2 GHz, both hyperthreading, with 6 GB RAM)

OS: Linux, RHEL Release 4, update 3, Kernel 2.6.9-34.0.1.0.11.ELsmp

Load: 225 calls/s, approximately 2800 SIP messages/s

CPU Average: 37.21%

Mem Usage: approx 650 - 1000 MB

Max GC Pause: 0.98551 s (from GC Viewer)

Avg GC Pause:  0.052 s

Comments
EVALUATION It appears highly likely (see comments section) that the problem that the customer is running into is the one discussed in CR 6336900 and addressed in the JVM via CR 6546278. We suggest that the customer upgrade to 6u4 or 5u14 where CR 6546278 has been fixed. If an upgrade is not feasible, the workaround documented in CR 6546278 should be used: -XX:+UseMembar. We are closing this as a duplicate of 6456278. Please let us know if the workaround/fix above do not address the issue. (We are fairly confident that it will.)
30-08-2007

EVALUATION Based on an analysis of the GC logs, we believe this may be related to 6546278 and related cluster of bugs. We have asked customer to run with -XX:+PrintGCApplicationStoppedTime -XX:+PrintGCApplicationConcurrentTime, and once we get the data we will update whether the data confirms our conjecture. Until that data is received we will mark this bug incomplete for process reasons related to the 6u5 release.
29-08-2007