Re: [BULK] Re: Some baseline tests on new hardware (was Re: [PATCH] xfs: optimise CIL insertion during transaction commit [RFC])

[Chris Mason <chris.mason@fusionio.com>]

Quoting Dave Chinner (2013-07-08 21:26:14)
> On Mon, Jul 08, 2013 at 09:15:33PM -0400, Chris Mason wrote:
> > Quoting Dave Chinner (2013-07-08 08:44:53)
> > > [cc fsdevel because after all the XFS stuff I did a some testing on
> > > mmotm w.r.t per-node LRU lock contention avoidance, and also some
> > > scalability tests against ext4 and btrfs for comparison on some new
> > > hardware. That bit ain't pretty. ]
> > > 
> > > And, well, the less said about btrfs unlinks the better:
> > > 
> > > +  37.14%  [kernel]  [k] _raw_spin_unlock_irqrestore
> > > +  33.18%  [kernel]  [k] __write_lock_failed
> > > +  17.96%  [kernel]  [k] __read_lock_failed
> > > +   1.35%  [kernel]  [k] _raw_spin_unlock_irq
> > > +   0.82%  [kernel]  [k] __do_softirq
> > > +   0.53%  [kernel]  [k] btrfs_tree_lock
> > > +   0.41%  [kernel]  [k] btrfs_tree_read_lock
> > > +   0.41%  [kernel]  [k] do_raw_read_lock
> > > +   0.39%  [kernel]  [k] do_raw_write_lock
> > > +   0.38%  [kernel]  [k] btrfs_clear_lock_blocking_rw
> > > +   0.37%  [kernel]  [k] free_extent_buffer
> > > +   0.36%  [kernel]  [k] btrfs_tree_read_unlock
> > > +   0.32%  [kernel]  [k] do_raw_write_unlock
> > > 
> > 
> > Hi Dave,
> > 
> > Thanks for doing these runs.  At least on Btrfs the best way to resolve
> > the tree locking today is to break things up into more subvolumes.
> 
> Sure, but you can't do that most workloads. Only on specialised
> workloads (e.g. hashed directory tree based object stores) is this
> really a viable option....

Yes and no.  It makes a huge difference even when you have 8 procs all
working on the same 8 subvolumes.  It's not perfect but it's all I
have ;)

> 
> > I've
> > got another run at the root lock contention in the queue after I get
> > the skiplists in place in a few other parts of the Btrfs code.
> 
> It will be interesting to see how these new structures play out ;)

The skiplists don't translate well to the tree roots, so I'll probably
have to do something different there.  But I'll get the onion peeled one
way or another.

-chris