It's Mac. I took a look at the issue you mentioned, just to be sure, and checked the test code there-- Mac doesn't seem to have the "can't let go of mmapped files" problem, at least.
I'll start digging deeper. --- A. Soroka The University of Virginia Library > On Jun 1, 2016, at 2:07 PM, Andy Seaborne <[email protected]> wrote: > > On 01/06/16 18:32, A. Soroka wrote: >> In this thread, which is about the testing for release of 3.1.0 >> >> https://lists.apache.org/thread.html/Z4t2lfn3qxqi571 >> >> a couple of people (myself, Osma) reported problems building on non-Linux >> platforms centered on problems with the enormous amount of space some tests >> seem to be taking on some systems. I'm still having that problem, and I'm >> wondering if anyone else has seen it or has any ideas about how to avoid it. >> It seems to pass through BlockMgrMapped.segmentAllocate, if that rings any >> bells for anyone. >> >> --- >> A. Soroka >> The University of Virginia Library >> > > MS Windows? > > On Windows, memory mapped files do not actually get deleted until the JVM, > exits [1]. > > So the test suite can't reuse a database by deleting it complete and reusing > the directory. > > A new, unique database is created each time (jena-tdb tests, > jena-jdbc-driver-tdb tests). > > Which means a lot of space. > > Maybe we ought to run the tests in "direct" mode on Windows. Thats' normal > old-fashioned file I/O, not this new fangled mapped stuff. > > If Mac, then I don't know. > > > Andy > > > > http://bugs.java.com/view_bug.do?bug_id=4715154
