On Mar 18 2000, andy huhn wrote:
> I thought linear was O(n)?
Indeed it is. But I think Uwe was referring to n searches,
each O(n).
> What are some more modern file systems that speed this up, and how
> do they do it? Do they use a hash for the structure of the
> directory?
On Sat, Mar 18, 2000 at 11:59:39AM -0500, andy huhn wrote:
> I thought linear was O(n)? What are some more modern file systems
linear is O(n).
But the programs are not reading the directories once, they open
every file once - n*(n/2).
> that speed this up, and how do they do it? Do they us
> Sent: Saturday, March 18, 2000 5:41 AM
> To: David E. Weekly
> Cc: [EMAIL PROTECTED]
> Subject: Re: Maildir Scalability Question
>
>
> On Sat, Mar 18, 2000 at 02:06:11AM -0800, David E. Weekly wrote:
>
> > my Maildir has 5086 messages in it to be precise. Recently (i.e.
On Sat, Mar 18, 2000 at 02:06:11AM -0800, David E. Weekly wrote:
> my Maildir has 5086 messages in it to be precise. Recently (i.e., in the
> last 500 messages or so) retreiving mail has become *painfully* slow.
> Looking at "top," I find that imapd is choking the CPU, taking 97% of the
> CPU ju
I'm running a Linux box as my mail server running qmail and a patched
version of the IMAP 4.5 server that supports Maildirs. I have a lot of
mail:
my Maildir has 5086 messages in it to be precise. Recently (i.e., in the
last 500 messages or so) retreiving mail has become *painfully* slow.
Looking