On Wed, Oct 31, 2018 at 2:37 AM Thomas Munro
<thomas.munro@enterprisedb.com> wrote:
> On Tue, Oct 30, 2018 at 4:15 AM Peter Geoghegan <pg@bowt.ie> wrote:
> > On Mon, Oct 29, 2018 at 2:11 PM Peter Geoghegan <pg@bowt.ie> wrote:
> > > This first line looks like it might be interesting:
> > >
> > > LOG: could not rmdir directory
> > > "base/pgsql_tmp/pgsql_tmp5088.0.sharedfileset": Directory not empty
> > > ERROR: could not determine size of temporary file "0"
> >
> > (Thomas Munro is CC'd here.)
> >
> > > I suppose that this could actually just be a result of the ERROR; the
> > > exact order isn't a reliable indicator of the sequence of events
> > > across processes (A useful log_line_prefix setting might clear this up
> > > if you collect the trace_sort output again).
> >
> > Hmm. So apparently Windows has a habit of setting an ENOTEMPTY errcode
> > when rmdir'ing a directory that somebody merely has a handle to. It
> > could just be that somebody has a Windows Explorer window open -- you
> > still get ENOTEMPTY [1]! Not sure if this is truly relevant to the
> > problem at hand, but it seems worth being aware of.
>
> We only try to remove the directory after removing everything in it,
> so that does seem to require a pretty strange explanation like that.
> I also wondered if a worker having a handle to an unlinked file (as
> permitted by the FILE_SHARE_DELETE flag we use) inside that directory
> could produce that effect. Perhaps RemovedDirectoryA[1] would be
> better than rmdir, since it "... marks a directory for deletion on
> close. Therefore, the directory is not removed until the last handle
> to the directory is closed." The documentation for rmdir[2] just says
> deprecated, and _rmdir[3] mentions ENOTEMPTY, but it says you get
> EACCES if someone has an open handle to the directory.
Is there a Windows developer who would like to experiment with this
stuff -- perhaps with a small standalone program that tries to unlink
a directory while it has another handle to the directory open -- and
make a recommendation? Otherwise I'll try to do this with AppVeyor.
While having the directory open in Explorer.exe might seem a bit
unusual for a production database, I bet there are other things that
periodically go around opening and reading our directories: backup
tools and SearchIndexer.exe presumably traverse the whole filesystem
from time to time. By my reading, RemovedDirectoryA() is probably the
right thing to defend against this, here and any other place we remove
directories.
--
Thomas Munro
http://www.enterprisedb.com
