Re: Large spool mboxes => slow?

> Date: Thu, 9 Mar 2006 15:21:27 +0000 (WET)
> From: Hugh Sasse <[email protected]>
> 
> On Thu, 9 Mar 2006, Gregory Hicks wrote:
> 
> > [Hugh Sasse wrote] 
> > > So, I have reached the conclusion that the best thing I can do to
> > > improve performance is to use the --enable-temp-drop-dir and point
> > > it to a different partition from /var/spool/mail so that disk seeks
> > > on both partitions may occur in parallel.  Does this sound like a
> > 
> > Point to different spindles on different I/O channels.  You have 
> > c0t0d0s6.  Use c1t0d0s6 for the temp-drop-dir.  At a minimum, point to 
> > different spindles.
> 
> OK, I think we only have c0t0d0s* on that machine, so I'm a bit stuck.
> I thought that might be the case.

Hugh:

If you have the Sun PCI SCSI board, that comes with two SCSI controllers...

In any case, try and get a second drive for the temp drop dir.

> > 
> > any POP3 daemon is going to have problem is the spool gets too large.
> > 
> > Also enable server mode.  enable caching of temp dir
> 
> Enabling server mode is not possible given some types of client (from 
> my reading of the docs).  I don't know all the clients people use, so 
> I must err on the side of caution.  Or is that paranoid in 2006?

I have to state that this is not the be-all, end-all of clients served,
but we have a quite diverse user population here.  I have found - via
the school of hard knocks - that the only client that the current
qpopper does NOT support is the old Z-Mail.  (I finally, 10 years after
the fact, got these users to upgrade to a more modern MUA...  Z-Mail
went out of business in 1996...)

Now I have to preface this with a "YMMV", but all 'modern' pop3 clients
should work with the current qpopper - including Outlook (even though I
would not wish this client on my worst enemy...)

> 
> Caching requires server mode....
> > 
> > > useful thing to do?  What if they are on different slices of the same
> > > disk -- would that make things worse (further for the heads to seek)? 
> > 
> > This WILL make things worse for just the reason you've stated.
> 
> Thanks.  I'm not au fait with disk internals, and thought that some 
> disks may have many heads, not just to read one cylinder at a time, 
> but possible several.  Access time is a market (selective) pressure
> on disks.

Well, they DO have multiple heads, but only one arm that these heads
are mounted on.  Best bet is to get multiple spindles.

Oh well...

Regards,
Gregory Hicks

> > 
>         Thank you
>         Hugh

-------------------------------------------------------------------
I am perfectly capable of learning from my mistakes.  I will surely
learn a great deal today.

"A democracy is a sheep and two wolves deciding on what to have for
lunch.  Freedom is a well armed sheep contesting the results of the
decision." - Benjamin Franklin

"The best we can hope for concerning the people at large is that they
be properly armed." --Alexander Hamilton