You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Rob Hartill <ha...@hyperreal.com> on 1995/10/06 21:18:15 UTC
Apache Patch: timeout vs abort (fwd)
ack sent
Forwarded message:
> From sabat@enterprise.Dts.Harris.COM Fri Oct 6 11:32:43 1995
> Date: Fri, 6 Oct 1995 11:30:42 -0700
> From: sabat@enterprise.Dts.Harris.COM (Steve Abatangle)
> Message-Id: <95...@velveeta.harris.com>
> To: apache-bugs@mail.apache.org
> Subject: Apache Patch: timeout vs abort
> Content-Length: 1401
>
> Apache .8.14 patch: timeout/abort
>
>
> This is a patch that lets http_main.c differentiate between client
> timeouts and client aborts. I whipped it up because
>
> a] it's useful information -- if you're getting a lot of timeouts,
> you may have network problems that need fixing. A large number
> of aborts might tell you your pages are too full of huge GIFs.
> But if you can't tell the difference between a timeout and an
> abort, you aren't able to spot these kinds of problems.
>
> b] NCSA has something similar in 1.4+.
>
>
> Steve Abatangle
> sabat@dts.harris.com
>
>
>
>
--- http_main.c 1995/10/6 11:16:45
230c230
< char errstr[MAX_STRING_LEN];
---
> char errstr[MAX_STRING_LEN],action[MAX_STRING_LEN];
244,248c244,259
<
< sprintf(errstr,"%s timed out for %s",
< timeout_name ? timeout_name : "request",
< current_conn->remote_name);
<
---
>
> if (sig==SIGALRM) { /* SIGALRM is a timeout */
> strcpy (action, "timed out");
> }
> else if (sig==SIGPIPE) { /* SIGPIPE is probably an abort */
> strcpy (action, "aborted");
> }
> else { /* extra wackiness just in case */
> strcpy (action, "stopped unidentifiably");
> }
>
> sprintf(errstr,"%s %s for %s",
> timeout_name ? timeout_name : "request",
> action,
> current_conn->remote_name);
>
Re: Apache Patch: timeout vs abort (fwd)
Posted by Brian Behlendorf <br...@organic.com>.
If the patch works as advertised, I consider this a good thing - and not
a functional enhancement (which would delay it until post-1.0)
Brian
On Fri, 6 Oct 1995, Rob Hartill wrote:
> ack sent
>
> Forwarded message:
> > From sabat@enterprise.Dts.Harris.COM Fri Oct 6 11:32:43 1995
> > Date: Fri, 6 Oct 1995 11:30:42 -0700
> > From: sabat@enterprise.Dts.Harris.COM (Steve Abatangle)
> > Message-Id: <95...@velveeta.harris.com>
> > To: apache-bugs@mail.apache.org
> > Subject: Apache Patch: timeout vs abort
> > Content-Length: 1401
> >
> > Apache .8.14 patch: timeout/abort
> >
> >
> > This is a patch that lets http_main.c differentiate between client
> > timeouts and client aborts. I whipped it up because
> >
> > a] it's useful information -- if you're getting a lot of timeouts,
> > you may have network problems that need fixing. A large number
> > of aborts might tell you your pages are too full of huge GIFs.
> > But if you can't tell the difference between a timeout and an
> > abort, you aren't able to spot these kinds of problems.
> >
> > b] NCSA has something similar in 1.4+.
> >
> >
> > Steve Abatangle
> > sabat@dts.harris.com
> >
> >
> >
> >
> --- http_main.c 1995/10/6 11:16:45
> 230c230
> < char errstr[MAX_STRING_LEN];
> ---
> > char errstr[MAX_STRING_LEN],action[MAX_STRING_LEN];
> 244,248c244,259
> <
> < sprintf(errstr,"%s timed out for %s",
> < timeout_name ? timeout_name : "request",
> < current_conn->remote_name);
> <
> ---
> >
> > if (sig==SIGALRM) { /* SIGALRM is a timeout */
> > strcpy (action, "timed out");
> > }
> > else if (sig==SIGPIPE) { /* SIGPIPE is probably an abort */
> > strcpy (action, "aborted");
> > }
> > else { /* extra wackiness just in case */
> > strcpy (action, "stopped unidentifiably");
> > }
> >
> > sprintf(errstr,"%s %s for %s",
> > timeout_name ? timeout_name : "request",
> > action,
> > current_conn->remote_name);
> >
>
>
>
>
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com brian@hyperreal.com http://www.[hyperreal,organic].com/