httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Jagielski <...@jaguNET.com>
Subject Re: error log messages in apache 1.3.27
Date Sun, 13 Jul 2003 22:29:37 GMT
Under FreeBSD, you'll get errno 22 if the socket is closed when
the setsocket call trys to set the TCP_NODELAY. It's unlikely
this is happening at the listener stage but likely right after
the accept of the request. If right after we accept the request
there is a period of time before we start reading it
that we set some socket params. If the socket is closed
at that point, the error will be logged.

IIRC, 1.3.28-dev has adjusted this behavior, as far as reducing
the severity of the log level entry and providing some sort
of hint on what happened.

Ian Holsman wrote:
> 
> Thanks Jim.
> I think it is FreeBSD.
> On Sunday, July 13, 2003, at 08:37  AM, Jim Jagielski wrote:
> 
> > With 1.3, we try to disable Nagle, by setting the socket to
> > TCP_NODELAY. the setsockopt() call fails, we continue, but log the
> > failure. We do this everytime we create a socket (which we
> > do when we create the listeners).
> >
> > What OS is this?
> >
> > Ian Holsman wrote:
> >>
> >> anyone seen/know what this is ?
> >>
> >> apparently he saw a similar thing on 1.3.20 when he got hacked
> >>
> >> Begin forwarded message:
> >>
> >>> From: <xxx@cnet.com>
> >>> Date: Sat Jul 12, 2003  5:28:27  PM US/Pacific
> >>> To: ianh@cnet.com
> >>> Subject: error log messages in apache 1.3.27
> >>>
> >>> Ian,
> >>>
> >>> Do you have any idea what the following error messages mean in
> >>> an apache 1.3.27 error log:
> >>>
> >>> [Wed Jul  9 20:09:21 2003] [warn] (22)Invalid argument: =20
> >>> setsockopt:=3D20
> >>> (TCP_NODELAY)
> >>> [Sat Jul 12 03:26:43 2003] [warn] (22)Invalid argument: =20
> >>> setsockopt:=3D20
> >>> (TCP_NODELAY)
> >>>
> >>> With the first there is no corresponding access log entry.  With
> >>> the second there is some activity at the same time (the server
> >>> is very lightly loaded).
> >>>
> >>> bye,
> >>> ke2
> >>>
> >> --
> >> Ian Holsman / 415-344-2608
> >> Performance Measurement & Analysis @ CNET Networks
> >>
> >> If everything seems under control, you're just not going fast enough. =
> =20
> >> =3D20=3D
> >>
> >> =3D97Mario Andretti
> >>
> >
> >
> > -- =20
> > =
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20
> > =3D=3D=3D=3D
> >    Jim Jagielski   [|]   jim@jaguNET.com   [|]   =20
> > http://www.jaguNET.com/
> >       "A society that will trade a little liberty for a little order
> >              will lose both and deserve neither" - T.Jefferson
> >
> --
> Ian Holsman / 415-344-2608
> Performance Measurement & Analysis @ CNET Networks
> 
> If everything seems under control, you're just not going fast enough.  =20=
> 
> =97Mario Andretti
> 


-- 
===========================================================================
   Jim Jagielski   [|]   jim@jaguNET.com   [|]   http://www.jaguNET.com/
      "A society that will trade a little liberty for a little order
             will lose both and deserve neither" - T.Jefferson

Mime
View raw message