RFR: JDK-8314272: Improve java.util.prefs.BackingStoreException: Couldn't get file lock.

[Matthias Baesken]

On Fri, 18 Aug 2023 12:14:35 GMT, Jaikiran Pai <jpai at openjdk.org> wrote:

> Hello Matthias, I think adding some diagnostics in this class would be a good thing. Like in the JBS issue that this PR links to, there's another JBS issue https://bugs.openjdk.org/browse/JDK-8304938 where we run into intermittent failures because of this exception.
> 
> However, before doing the changes being proposed in this PR, would it be better to do those (minimal) changes in a private build and run those tests and see if these proposed error number or error message diagnostics do provide the needed detail to narrow down this issue? That might provide insight on what else (if anything) would need to be included in that diagnostic message.
> 
> In the meantime, I'll run your proposed change which shows the error number, just to see if that tells us what's going on with these tests, when they fail.

the enhanced output in a failing sun/tools/jhsdb/JStackStressTest.java  is 

java.util.prefs.BackingStoreException: Couldn't get file lock. errno is 11 mode is nonshared

errno  11 seems to be EAGAIN. Unfortunately  even the enhanced message does not tell me if it is a failure of open or fcntl in the native layer , that would be good to know.  
We see this issue mostly on a particular Linux machine, probably I should ask the admins what's 'special' about this machine ....

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

PR Comment: https://git.openjdk.org/jdk/pull/15308#issuecomment-1683850190