I believe it can happen with any of the tests that create an instance of a DrillBit and Client. A specific example would be ParquetRecordReaderTest.testMultipleRowGroupsAndReads2. I've been running into it with my options tests as well though, so it doesn't appear to have anything to do with parquet. I have included the full logs as exported with lilith as attachments.
On Fri, Mar 28, 2014 at 4:39 PM, Aditya <[email protected]> wrote: > Any test in particular or just any random test? > > > On Fri, Mar 28, 2014 at 2:37 PM, Jason Altekruse > <[email protected]>wrote: > > > Hello Drillers, > > > > I have been running into an issue with running tests in intellij. There > > does not seem to be a consistent cause, but it manifests in consistent > > error occurring. It can been seen in lilith, but is not caught by Drill > and > > causes queries to spin endlessly. The actual message leads to a line of > > code in our forked PooledByteBufAllocatorL which has been replaced by an > > explicit throwing of an exception (I do not believe we are supposed to be > > getting down this code path in regular execution but there isn't a note > in > > the code about why the exception is being thrown its just an > > UnsupportedOperatorException). > > > > The best I can guess is it is related to left over remnants from previous > > tests (it seems to happen less frequently if I leave a few seconds > between > > test runs). > > > > Has anyone else experienced this? > > > > -Jason > > > > >
