JDK-6721619 : Java Web Start 1.6.0 shows intermittent https communication breakdown
  • Type: Bug
  • Component: core-libs
  • Sub-Component: java.net
  • Affected Version: 6
  • Priority: P2
  • Status: Closed
  • Resolution: Not an Issue
  • OS: windows_xp
  • CPU: x86
  • Submitted: 2008-07-02
  • Updated: 2011-12-15
  • Resolved: 2010-03-22
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.
6u12Fixed 7Resolved
Related Reports
Duplicate :  
Java Web Start application started on Java Web Start 1.6.0 intermittently 
fails to download resource files over https. The problem can be reproduced
with 1.6.0_03, 1.6.0_06 and 1.6.0_10 b25. 
Everything is fine when using Java Web Start 1.5.0 on JRE 1.6.0 in exactly
the same enviroment.
So, Java Web Start 1.6.0 is less robust/reliable wrt https communication 
than Java Web Start 1.5.0.

EVALUATION Since the changes for 6882654 (Remove dependency on java.util.concurrent from KeepAlive implementaion) this issue is no longer a problem in JDK7. CR 6882654 changes the synchronization where KeepAliveStreams are put on the queue and does not suffer from a race where the comsumer may remove the streams and clean them before the producer has finished setting the queuedForCleanup flag.

EVALUATION There is a race condition in the KeepAliveStream implementation that may allow a KeepAliveStream instance to be cleaned up by the KeepAliveStreamCleaner before it has its queuedForCleanup flag set. This may lead to the Http(s)Client instance being put in the KeepAliveCache twice. This may result in the KeepAliveCache closing the underlying socket while the jaws implementation is downloading on the same socket. The solution is to synchronize on the KeepAliveStream instance while offering it to the cleaner queue as well as setting its queuedForCleanup flag. This will block the clearner from performing any action on the KeepAliveStream instance before it gets its queuedForCleanup flag set.

WORK AROUND use -Dhttp.keepAlive=false for Java Web Start 1.6