Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 1 Apr 2010 14:00:36 -0500
From:      Peter Steele <psteele@maxiscale.com>
To:        Steve Polyack <korvus@comcast.net>
Cc:        "freebsd-questions@freebsd.org" <freebsd-questions@freebsd.org>
Subject:   RE: Testing ethernet interface status
Message-ID:  <7B9397B189EB6E46A5EE7B4C8A4BB7CB3B65FC97@MBX03.exg5.exghost.com>
In-Reply-To: <4BB4EC44.2070804@comcast.net>
References:  <7B9397B189EB6E46A5EE7B4C8A4BB7CB3B65FC5D@MBX03.exg5.exghost.com> <4BB4EC44.2070804@comcast.net>

next in thread | previous in thread | raw e-mail | index | archive | help
>I was going to suggest that you look at the ifconfig(8) source code, but t=
hen I did so myself - it looks like you're doing it pretty much exactly how=
 they are.  I've never >noticed ifconfig(8) returning an incorrect value, n=
ot to say it's not possible.
>
>Are you sure that nothing is causing interface state resets? i.e.=20
>mismatched duplex/speed settings between the FreeBSD machine and the switc=
h?  Have you checked dmesg(8) for logs of interface state changes? =20
>You can also check the output of 'netstat -i' to check for interface error=
s.

I should have added that when our own monitoring code flags one of these fa=
lse positives, there is no entry in /var/log/messages indicating that the n=
ic has gone offine. I added a second call to confirm that indeed the interf=
ace is offline, and this second check seems to have largely solved the prob=
lem, but we have seen a case where even two consecutive checks return false=
 positives. Maybe we need three tests?




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7B9397B189EB6E46A5EE7B4C8A4BB7CB3B65FC97>