procmail
[Top] [All Lists]

Re: size of maildir vs. mbox

2001-11-06 19:53:33
I'm no expert, but I think the size increase will be a function of the
average (or median?) message size and the inode (block) size of the
filesystem, namely:
  (inode size)/(2 * message size)

This is just the typical "wasted" space on the filesystem, assuming the
files have roughly random sizes.  In other words, on average, each message
will waste half an inode.  However, if the messages tend to be less than
half the inode size, the waste will be even more.

Looking at my own mailbox, my average message size is about five kilobytes
while my inode size is four kilobytes, so the increase in size -- if and
when I ever get maildir figured out -- will be about 40%.

Naturally, if you can put the maildirs on a filesystem with smaller inodes
(2048 or 1024 bytes), you can reduce this overhead.  The other way to
reduce the overhead is to make the mails bigger . . . but that would
really defeat the purpose, wouldn't it?  ;-)

James
LP^>

On Tue, Nov 06, 2001 at 01:56:49PM -0500, Paul Chvostek wrote:

Can anybody advise how much larger a mail spool will be, using maildirs
rather than mbox format mailboxes?  I could probably write a script to
come up with exact numbers, but if someone's done the switch already and
noted an X percent increase in disk usage, I'd love to know what X is.
I can't seem to find any discussion of this in MTA FAQs.

Thanks.


-- 
  Paul Chvostek                                             
<paul(_at_)it(_dot_)ca>
  Operations / Development / Abuse / Whatever       vox: +1 416 598-0000
  IT Canada                                            http://www.it.ca/

_______________________________________________
procmail mailing list
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail
_______________________________________________
procmail mailing list
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail

<Prev in Thread] Current Thread [Next in Thread>