mutt + procmail + nfs...

2002-07-31 Thread kevin lyda

ok, i have a mail server that also exports out /home.  it has several
raid disks is connected to an ups and is generally noisy so it lives in
my utility room.  my workstation lives in my office and it mounts /home
from the mail server.

sendmail invokes procmail to deliver my mail on the mail server.
i'm thinking of having procmail deliver directly to ~/Mail/inbox.
that way i could run mutt on my workstation.

however, this is the dreaded mailbox on nfs issue.  do procmail and mutt
play nice on nfs?

kevin

-- 
[EMAIL PROTECTED] that a believer is happier than a skeptic is no more to
fork()'ed on 37058400the point than the fact that a drunken man is happier
meatspace place: home   than a sober one. the happiness of credulity is a
http://ie.suberic.net/~kevin   cheap  dangerous quality -- g.b. shaw



msg30003/pgp0.pgp
Description: PGP signature


Re: mutt + procmail + nfs...

2002-07-31 Thread Ralf Hildebrandt

On Wed, Jul 31, 2002 at 05:45:32PM +0100, kevin lyda wrote:
 ok, i have a mail server that also exports out /home.  it has several
 raid disks is connected to an ups and is generally noisy so it lives in
 my utility room.  my workstation lives in my office and it mounts /home
 from the mail server.
 
 sendmail invokes procmail to deliver my mail on the mail server.
 i'm thinking of having procmail deliver directly to ~/Mail/inbox.
 that way i could run mutt on my workstation.
 
 however, this is the dreaded mailbox on nfs issue.  do procmail and mutt
 play nice on nfs?

If you use Maildir, yes.

-- 
Ralf Hildebrandt (Im Auftrag des Referat V A)   [EMAIL PROTECTED]
Charite Campus Virchow-Klinikum Tel.  +49 (0)30-450 570-155
Referat V A - Kommunikationsnetze - Fax.  +49 (0)30-450 570-916
This is the first age that's paid much attention to the future, which is
a little ironic since we may not have one. -- Arthur Clarke




Re: mutt + procmail + nfs...

2002-07-31 Thread darren chamberlain

* kevin lyda [EMAIL PROTECTED] [2002-07-31 12:44]:
 however, this is the dreaded mailbox on nfs issue.  do procmail and
 mutt play nice on nfs?

Mutt and procmail both support Maildir natively. See
http://cr.yp.to/proto/maildir.html.

(darren)

-- 
Patriotism is the last resource of scoundrels.
-- Samuel Johnson



Re: mutt + procmail + nfs...

2002-07-31 Thread Brad Knowles

On Wed, Jul 31, 2002 at 06:46:55PM +0200, Ralf Hildebrandt wrote:

  however, this is the dreaded mailbox on nfs issue.  do procmail and mutt
  play nice on nfs?
 
 If you use Maildir, yes.

Maildir has the problem of trading NFS locks for excess synchronous
meta-data operations, and synchronous meta-data operations are
something not dealt with well by most NFS servers (at least, anything
short of Auspex, EMC, or NetApp).

There are alternative solutions.


For POP3-based systems, Nick Christenson led the way in his
Earthlink paper, _A Highly Scalable Electronic Mail Service Using
Open Systems_, which you can find at
http://www.jetcafe.org/~npc/doc/mail_arch.html.  IIRC, they used file
creation tricks (just like Maildir does), but instead of breaking out
each message into a separate file, they used the v7 mbox file
format.

You still have some synchronous meta-data issues to deal with, but
not nearly on the scale of Maildir, and even those can be at least
partially addressed (start with
http://www.shub-internet.org/brad/papers/dihses/lisa2000/sld065.htm
of my invited talk _The Design and Implementation of Highly Scalable
E-mail Systems_, which I presented at LISA 2000).


For IMAP-based systems, you would need to use the message store
engine from MessagingDirect (they hacked Cyrus so as to pull the
locking out of the filesystem and into an NFS-friendly database).
Alternatively, you would need to hack Cyrus to pull the file locking
out into an NFS-friendly database.


You may also be interested to read the book _Sendmail Performance
Tuning_ by Nick Christenson -- to be published on September 20, I
believe.  While the title of the book is about sendmail, I believe that
it covers more than just this one program.

Disclosure: I was a technical reviewer of this book and Nick was my
co-author for the talk I gave at LISA 2000, so I may be a bit biased.
;-)

-- 
Brad Knowles [EMAIL PROTECTED] Senior Consultant for Snow bv
ASM Lithography bv  + ITMS Unix Competence Center  + Room 02.D.2085
De Run 1110 + 5503 LA Veldhoven, NL + GSM +31 654 344 596