httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chuck Murcko <ch...@telebase.com>
Subject Re: mod_status again
Date Thu, 02 May 1996 21:32:40 GMT
Jim Jagielski liltingly intones:
> 
> Unfortunately, that really doesn't provide a lot of info. The reason I
> implemented it (the 2 copies) was to account for things like how many
> the _current_ process has handled since it's been alive. This is different
> from the total number for that _slot_ in the scoreboard file. I think
> the current implementation really reduces the amount of info that's
> useful since it tracks the _current_ use and not the culmulative (sp)
> use for that child, and that's what we are most interested in. The
> info about the last connection is nice, but not as useful as a inside
> look on the entire child activity.
> 
> If the group thinks the *last* info is useful, I would prefer adding
> 2 more longs to the scoreboard to hold _that_ info as well... But this
> is a dangerous road to start down (making the scoreboard bigger) unless
> we really think the info is useful. Personally, if I had to choose, I
> would prefer the the way it was over this... it gives me more
> insight.
> 
It looks like the current slot total is cleared when a child dies, and starts
incrementing again when the slot is reused. I actually had added 2 more longs
to the scoreboard in earlier testing, but as I wrote earlier, I am used to
running with MaxRequestsPerChild at 0, so my child and slot totals are always
the same, which is also not so informative. Reverting to the original
http_main.c restores the original behavior.

I had considered this a stopgap until actual snapshots of current connect
and byte counts were in place. This is transient data, and needn't take up
space in the scoreboard. The last info is, as you say, less useful.
Average time to service a request is something I've been toying with
here for awhile, and might also be useful. That wouldn't be in the score-
board either, so it shouldn't be a big problem. As it is, either way the
first number of Accesses and the B1 value is displayed, the scoreboard
hasn't changed size. Nor does the status module make any more passes
through the scoreboard than it used to.

chuck
Chuck Murcko	N2K Inc.	Wayne PA	chuck@telebase.com
And now, on a lighter note:
It wasn't that she had a rose in her teeth, exactly.  It was more like
the rose and the teeth were in the same glass.

Mime
View raw message