Here is a second one.
--
:wq Claudio
--- Begin Message ---
I've been doing some non-scientific performance tests
on my latest QmailLDAP/Controls patch...
These times are for receiving 1000 mails sent from
remote to local. Does not time the actual time from
receive to delivery...
It's bad, but it's not THAT bad. You have to receive _A LOT_
(millions?) of mails every hour for this to be a concern...
==============================================
QMail original: 44s (SLAPDs: n/a)
QMailLDAP: 47s (SLAPDs: 207)
QMailLDAP/Controls (1): 403s (6m43s) (SLAPDs: 289)
QMailLDAP/Controls (2): 122s (2m02s) (SLAPDs: 1013)
QMailLDAP/Controls (3): 144s (2m24s) (SLAPDs: 1014)
Note 1:
QMailLDAP/Controls (1): bind/unbind for each control value
QMailLDAP/Controls (2): no unbind what so ever. Global 'LDAP *',
which isn't free'd etc.
QMailLDAP/Controls (3): Much better 'hookup' to existing QmailLDAP
code...
Note 2:
* Initial SLAPDs running = 4 (sometimes 8!?).
* SLAPd's not dying!
Delivering 2000 mails (1000 mails from two hosts) takes
roughly 5 minutes, and that's still pretty ok... No change
in number of slapd's running (probably a sysctl thing), so
those 1000 slapd's isn't DEAD, they're just not shutting
down as the're supposed to...
What's weird is that CTRL(2) is _faster_ than CTRL(3). I had
expected the high amount of slapd's in CTRL(2), but NOT in
CTRL(3)! It should be possible to be more resourcefull with
LDAP connections, but i'm just not seeing it...
--
critical munitions security congress North Korea ammunition Mossad
smuggle tritium Peking Waco, Texas NSA counter-intelligence Ft. Bragg
genetic
[See http://www.aclu.org/echelonwatch/index.html for more about this]
--- End Message ---
