-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 10 Mar 00, at 12:36, Anand Buddhdev wrote:
> Are you saying that spending a few more resources checking up valid
> usernames is better than accepting possibly large emails, and then
> attempting to bounce them, and spending bandwidth, time and
> qmail-remote slots, which would be better used for genuine outgoing
> emails?
Yes, in a way. ("Don't speculate, profile." And I _am_ speculating.
Aren't you as well?) How many _large_mails_ to nonexistent users
do you receive? The ones I receive are around 5kB; bandwidth hit is
negligible. More importantly, in case of overloaded box, most of the
actions neccessary to generate the bounce is rescheduled for
later. (Simple queue the mail, and "wait" for the load to drop - I
mean, wait till qmail-send is free to do something about the
message.)
> I think the problem of slow lookups in large databases has gone away
> with CDB.
1. Can you really describe all your users in CDB? Even all the
user-ext stuff? Do you disallow user access to "ezmlm-make"? Or
do you force CDB rebuild after each .qmail-* editing?
2. CDB lookup may be much faster than an ordinary database
lookup, but it still requires disk activity; I wouln't assume that the
whole CDB fits into the memory.
3. Of course, YMMV :-)
-----BEGIN PGP SIGNATURE-----
Version: PGP 6.0.2 -- QDPGP 2.60
Comment: http://community.wow.net/grt/qdpgp.html
iQA/AwUBOMi2HlMwP8g7qbw/EQKwrwCgmJlP2oX3T1E/xJ9PEzsgM+uIRrsAniuK
QZSeSbmf5lLqFYN1TKJoIk9s
=pxKW
-----END PGP SIGNATURE-----
--
Petr Novotny, ANTEK CS
[EMAIL PROTECTED]
http://www.antek.cz
PGP key ID: 0x3BA9BC3F
-- Don't you know there ain't no devil there's just God when he's drunk.
[Tom Waits]
