[PATCH] TDB traverse lock changes for massive AD DC perf improvement

[Andrew Bartlett]

On Wed, 2017-04-05 at 11:51 +0200, Stefan Metzmacher via samba-
technical wrote:
> Am 05.04.2017 um 01:50 schrieb Andrew Bartlett via samba-technical:
> > On Wed, 2017-04-05 at 10:42 +1200, Andrew Bartlett wrote:
> > > On Sat, 2017-04-01 at 18:44 +1300, Andrew Bartlett via samba-
> > > technical
> > > wrote:
> > > > 
> > > > Aside from a good delay so others can comment, what would you
> > > > recommend
> > > > I do next towards progressing this patch?
> > > 
> > > Can I please get a good eye over and a team review for this
> > > patch?
> > > 
> > > Thanks,
> > > 
> > > Andrew Bartlett
> > 
> > The attached patch takes on board the 80 column feedback (I've also
> > found an emacs mode that is helpful but unobtrusive). 
> > 
> > I hope this meets with your approval, as it makes a massive
> > difference
> > to our Samba AD DC performance.
> > 
> > Please review.  If reviewed, I'll push with a patch that adds new
> > performance tests that I'm keen to get in. 
> 
> I'm wondering about all the readonly checks in
> _tdb_transaction_prepare_commit(),
> we already handle that in _tdb_transaction_start().
> 
> I'm a bit nervous about the solaris10 problem.

I am to.  I only got game to formally propose it when Jeremy
essentially proclaimed it dead :-)

> Ralph gave me access to a solaris 10 box, I'll try to reproduce the
> original
> problem there and check if we deadlock with the current code.

For the record (I included the git hashes of the history because it was
non-trivial to dig up), do you know or remember what the original
problem was?

Thanks,

Andrew Bartlett

-- 
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba