[Xapian-discuss] Update under a large database is slow

[James Aylett]

On Fri, Sep 07, 2007 at 05:23:28PM +0800, Gea-Suan Lin wrote:

> Running FreeBSD 7.0-CURRENT, using Gigabit Ethernet to access NetApp
> with NFSv3.

This *probably* isn't hitting you, but we've had problems with NFSv3
against some NetApp heads. It depends a lot on your loading profile,
the amount of data you're playing with, which heads and probably which
firmware you're using, etc. etc. etc.. It's very obvious if you hit
it: your NetApp CPU utilisation curves upwards really badly beyond a
certain scale, and then you start getting enormous asvc_t / IO wait
times, and everything grinds to a halt. If you have a good NFSv4
implementation, this works much better (for various reasons).

Or you could upgrade to bigger NetApp heads, which is one of the
things we had to do around that time ;-)

J

-- 
/--------------------------------------------------------------------------\
  James Aylett                                                  xapian.org
  james at tartarus.org                               uncertaintydivision.org