Version ages are fair enough, I'll get mysql 5.0 installed and run that. I just got done putting postgres 8.2 in place, as a matter of fact, and it delivered the 1000 messages in 9 seconds ;) Seemed to perform a little better than 8.1.4, although on a scale that's harder to find a worthwhile metric to show.
It's going to take a fair amount of effort to test pop/imap simulating multiple users concurrently, though. So, that part may take a while I'm afraid; unfortunately this project isn't my highest priority at the moment, I just find it fun and interesting. Out of curiosity, do you all have any requests or suggestions as to how many concurrent sessions you'd like to see? On Thu, 07 Dec 2006 10:40:19 -0500 Matthew O'Connor <[email protected]> wrote: > Lars Kneschke wrote: > > Justin McAleer <[EMAIL PROTECTED]> schrieb: > >> Mysql: 95 seconds to deliver all 1000 messages. Both > cores > >> on the DB server were effectively peaked during > delivery. > >> > >> Postgres: 10 seconds to deliver all 1000 messages. > DBMail > >> was really close to being able to deliver as fast as > >> postfix could queue to local disk (within a second or > two > >> for 1000th message). The cores on the DB server looked > to > >> average around 45%/30% usage during delivery. > > > > MySQL 4.1 is already very old. And MySQL 5.0 is much > more faster then 4.1. > > Do you think, you can test with MySQL 5.0 again? You > can donwload ready to > > go packages from mysql.com. > > Agreed, PG 8.1.4 is about a year old and was the most > recent release until yesterday when 8.2 came out, so we > probably aren't comparing apples to apples with that old > of a MySQL version. But while you are at it, any chance > you can run PG 8.2 through the same test? > > > On the other side, it would be more interesting for me, > how fast we can > > read from the databases. Inserting messages is not > something which needs to > > be fast. It's more interesting for me how fast we can > read messages from > > the imap server. > > Agreed also, I think this is where most of the load comes > from using DBMail especially IMAP, so this would be the > more compelling bench mark. > > Matt > _______________________________________________ > DBmail mailing list > [email protected] > https://mailman.fastxs.nl/mailman/listinfo/dbmail >
