Are you using hdfs ? If so, can you check namenode audit log ?
Cheers On Wed, Aug 26, 2015 at 12:01 PM, donmai <[email protected]> wrote: > Can't seem to find anything in the logs that even matches /foo/foo being > created. Any ideas as to where I can start looking in the codebase to > figure out exactly what's going on? > > On Wed, Aug 26, 2015 at 1:20 PM, donmai <[email protected]> wrote: > > > DEBUG is on, trying to look through the logs again. Thanks! > > > > On Wed, Aug 26, 2015 at 12:59 PM, Ted Yu <[email protected]> wrote: > > > >> Did you enable DEBUG logging ? > >> > >> Can you find the logs for one such occurrence and pastebin relevant > >> portion > >> ? > >> > >> Thanks > >> > >> On Wed, Aug 26, 2015 at 9:58 AM, donmai <[email protected]> wrote: > >> > >> > Hi, > >> > > >> > Occasionally when I run restore_snapshot on HBase 0.98.10, it appears > >> that > >> > the table directory structure created by the restore_snapshot command > is > >> > not correct: > >> > > >> > rootdir/data/default/foo > >> > > >> > Is what it should be, but I end up with > >> > > >> > rootdir/data/default/foo/foo > >> > > >> > The extra foo should not be there, so the master gets stuck in a loop > >> > trying to figure out why there isn't a > .tabledesc/.tableinfo.0000000001 > >> in > >> > rootdir/data/default/foo (it's in rootdir/data/default/foo/foo). > Moving > >> > everything from rootdir/data/default/foo/foo to > >> rootdir/data/default/foo/ > >> > unblocks HBase master and allows it to proceed. > >> > > >> > This doesn't always happen, so my question is: why does it happen? The > >> logs > >> > don't seem to be showing anything. > >> > > >> > Thanks! > >> > > >> > > > > >
