Yes, once it is finished everything is back to normal, even if I don't do mvn clean. It seems to be like you say, that ecryptfs garbage collection is lagging a bit behind all the fast transactions.
Perhaps adding something similar to this disk space check? https://github.com/wf4ever/robundle/blob/master/src/test/java/org/purl/wf4ever/robundle/fs/MemoryEfficiencyIT.java#L117 It can use ignore instead of failing.. if only I remembered the correct @Junit thing for that kind of pre-flight checks. On 29 October 2014 11:38, Andy Seaborne <a...@apache.org> wrote: > On 28/10/14 20:31, Stian Soiland-Reyes wrote: >> >> Hi, >> >> I am struggling to run "mvn install" on jena-tdb locally - as I seem >> to only have 3.5 GB free disk space. During the tests, this runs down >> towards 0. >> >> This is on Linux Mint 17 x64 with ecryptfs encrypted home directory. > > > TDB databases when created are sparse files ... space is allocated but not > used. I would not be surprised if ecryptfs blocked that possibility. > > Create a database ("tdbquery --loc DBX 'ASK{}') and run "du -sh" on the > directory. It is about 150-200K but ~192M if not OS sparse files. > > On Linux, databases are deleted and recreated. Again, ecryptfs may > interfere with that, delaying the file recycling into the OS as it clears up > - I don't use it so I am just speculating here. > > Is the space cleaned out after the test suite runs successfully? > >> This seems to be eaten up while running TestTransactionTDB. Is it >> keeping lots of mmap-ed temporary files open? > > > They should not be open. On linux they should be deleted and released to > the OS. > >> Is there a way to reduce this disk space requirement (which sounds >> more like an integration test to me) - or should I simply start >> tidying my /home ? :) > > > Its the tests of the transaction mechanism on disk. If you are not working > in that area, you could skip that test file. > > (it will also run the tests on an in-memory dataset so you are completely > running testless). > > And tidy /home anyway! Do I sense some guilt here? :-) > > Andy >> >> >> > > > > See also : JENA-755 -- Stian Soiland-Reyes, myGrid team School of Computer Science The University of Manchester http://soiland-reyes.com/stian/work/ http://orcid.org/0000-0001-9842-9718
