JDK-7017975 : Two memory leak tests fail just on solaris10 sun4v against jdk5 (is not particular update related)
Type:Bug
Component:core-libs
Sub-Component:java.net
Affected Version:5.0u28
Priority:P3
Status:Resolved
Resolution:Fixed
OS:solaris_10
CPU:sparc
Submitted:2011-02-08
Updated:2011-05-17
Resolved:2011-05-17
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.
EVALUATION
Might be this issue potantially has more general scope that some java apps may consume more memory be executed on Niagara against the others platforms because of the native heap "grows too fast"?
10-02-2011
EVALUATION
Niagara boxes have different heap and GC characterisics (22 GC threads for starters) . I've noticed the java heap growing quickly even after the 30-40 second wait to obtain "working heap" stats.
Limiting the testcase to a small maximum value for jav ahealp lets the process heap size become more stable. Any large growth in process heap can be attributed to native memory leak as a result. Will update both testcase with a -Xmx16m flag.
09-02-2011
EVALUATION
This test is not in JDK6 or JDK7 (adding 6-na and 7-na keywords).
The test sleeps for 10 seconds waiting for the java process to stabilize before capturing the initial memory usage. This doesn't appear to be long enough on sun4v systems. After running the test several times the memory usage appears to stabilize and not grow indicating that the leak bug is not back. More likely an issue with the testcase itself.