On Mon, Jan 13, 2003 at 08:03:20AM -0500, Sam Varshavchik wrote: > Brian Candler writes: > > > The benefit of this is that in the normal case where a client downloads and > > deletes all messages, no courierpop3dsizelist file needs to be written at > > all - thus there is no overhead in terms of disk space, inodes, and disk > > operations to maintain it. > > The overhead here is a single open(), write(), and close() system call. > Nothing to lose sleep over.
We're talking a Netapp with a few hundred thousand mailboxes and growing. If I can save one inode and one data block per mailbox, and one open-read-close open-write-close-rename set of NFS operations per login, that's a bonus to me. > > As an additional optimisation it could be not written out if no changes have > > been made (i.e. for repeated polls of a mailbox when no new items have > > arrived) > > And on the flip side, if a mailbox with new mail is opened twice, both > sessions will have to calculate the message sizes. Certainly. It's a question of which happens more often: two people simultaneously opening the same mailbox, or one person polling their mailbox more frequently than they receive mail. I'd be pretty confident that the second case is more prevalent. Regards, Brian. ------------------------------------------------------- This SF.NET email is sponsored by: FREE SSL Guide from Thawte are you planning your Web Server Security? Click here to get a FREE Thawte SSL guide and find the answers to all your SSL security issues. http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en _______________________________________________ courier-users mailing list [EMAIL PROTECTED] Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
