ZFS - abysmal performance with samba since upgrade to 8.2-RELEASE

[Jeremy Chadwick]

On Thu, Feb 24, 2011 at 07:23:54PM +0100, Christer Solskogen wrote:
> On Thu, Feb 24, 2011 at 8:55 AM, Jeremy Chadwick
> <freebsd at jdc.parodius.com> wrote:
> 
> >
> >  # Set TXG write limit to a lower threshold.  This helps "level out"
> >  # the throughput rate (see "zpool iostat").  A value of 256MB works well
> >  # for systems with 4GB of RAM, while 1GB works well for us w/ 8GB on
> >  # disks which have 64MB cache.
> >  vfs.zfs.txg.write_limit_override=1073741824
> >
> >
> 
> Sorry if you have said this before, but could you elaborate a bit
> about this number? For instance, how much does the cache on the disk
> has to say.
> In my case: 3x1.5TB raidz with WD15EADS-00R6B0 which has 32MB cache
> and 12GB memory. What would you recommend and why.

There's no real way to provide an in-depth analysis of this number; that
is to say, hard disk parameters (RPM, cache, and overall performance of
the drive (highly dependent upon on-disk firmware)) ultimately dictates
what's "best" for this number.  I also imagine number of disks plays a
role as well  This is why I advocate not messing with it unless you want
to try and "level out" throughput.

The value itself is literally 1024*1024*1024 (1GB).  Don't think this is
some magic number; it's just what I came up with.  You can search the
FreeBSD lists for references to the variable itself and find other
people advocating values like ~33MByte, etc..  All depends on how you
want the system to behave.

I don't particularly like watching the system behave like I described
(you snipped that portion of my text), which is why I use this variable.

-- 
| Jeremy Chadwick                                   jdc at parodius.com |
| Parodius Networking                       http://www.parodius.com/ |
| UNIX Systems Administrator                  Mountain View, CA, USA |
| Making life hard for others since 1977.               PGP 4BD6C0CB |