Mailing List Archive

> Vas ees das? (in the error_log)
>
> [Tue Aug 8 05:17:10 1995] Memory hog alert: allocated 139880 bytes for GET /HTTP/1.0
> [Tue Aug 8 05:17:21 1995] Memory hog alert: allocated 131048 bytes for GET /docs/images/apache_logo.gif HTTP/1.0
> [Tue Aug 8 05:17:30 1995] Memory hog alert: allocated 131200 bytes for GET /images/orange_ball.gif HTTP/1.0
> [Tue Aug 8 05:17:36 1995] Memory hog alert: allocated 131048 bytes for GET /images/apache_pb.gif HTTP/1.0
>
> Brian

Oh shit, my log's full of these too.

A quick look at the code (where the message is generated) doesn't
imply anything is broken, but is the message meaningful ?.. it points
some blame at innocent URLs e.g. apache_pb.gif above.

Shouldn't the message point blame at a child process that is growing, and
then mark that child to be zapped ASAP.


rob

Date: Wed, 9 Aug 1995 18:15:19 -0700 (PDT)
From: Brian Behlendorf <brian@organic.com>
Precedence: bulk
Reply-To: new-httpd@hyperreal.com


Vas ees das? (in the error_log)

[ Memory hog alerts... ]

I tossed that into a *very* early Shambhala when I was trying to
figure out which requests were blowing up server processes to nearly a
megabyte. (They were directory listings for long directories ---
about a thousand files --- which were generating about a thousand
sub_requests --- that's when the sub-pool mechanism came in, so that
memory could get freed up before the request was done).

Around now, it can be safely flushed.

rst