Hi,

I've been investigating the speed differences between
analog v2.11 and v3.x and have discovered that v3
seems to take approx. 3 times longer for a stats run
than v2.

For a large run (a couple of days worth) this can
make quite a difference, even with a DNS cache.

truss shows the analog process sleeping in door_call,
so I'm tempted to blame Solaris' nscd (which is over
70Mb now).

Does anyone have any experience or recommendations
for dealing with this; will things work better if
I just kill nscd or is there alternative tuning that
I can do to try to improve these times?

This wouldn't explain the discrepancy between the v3
and v2 runtimes, but it's not impossible that the v2
binary came from a machine which didn't know about
nscd.

Also: does anyone know why the below config file won't
read from compressed log files?  I get a "Broken Pipe",
though running a similar command at the shell seems
to work fine.

Cheers,
Matthew.


Here's an example config file.

LOGFILE /www/logs/imm_access_log*
OUTFILE /www/stats/imm/stats-19990715.txt
LOGO none
HOSTNAME imm
HOSTURL http://www.inmotionmagazine.com
OUTPUT ASCII
GRAPHICAL OFF
DAILY ON
DOMAIN ON
FROM -00-00-07
TO -00-00-01

DNSFILE /www/logs/dns.txt-1
DNSFRESHHOURS 672

UNCOMPRESS *.gz,*.Z "/usr/local/bin/gzip -cd"

------------------------------------------------------------------------
This is the analog-help mailing list. To unsubscribe from this
mailing list, send mail to [EMAIL PROTECTED]
with "unsubscribe analog-help" in the main BODY OF THE MESSAGE.
List archived at http://www.mail-archive.com/[email protected]/
------------------------------------------------------------------------

Reply via email to