Date: Sun, 11 Oct 2015 22:48:03 -0700 From: Doug Hardie <bc979@lafn.org> To: Terje Elde <terje@elde.net> Cc: FreeBSD - <freebsd-questions@freebsd.org> Subject: Re: Ethernet Errors Message-ID: <28F04ECC-AF7C-4089-808B-51807CD110A9@lafn.org> In-Reply-To: <9D927A93-DB81-4423-B556-31120DA3136B@elde.net> References: <C04A0076-C7CF-4D42-9BBD-546363CD3E19@lafn.org> <9D927A93-DB81-4423-B556-31120DA3136B@elde.net>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 11 October 2015, at 06:39, Terje Elde <terje@elde.net> wrote: >=20 > On 11. okt. 2015, at 12:14, Doug Hardie <bc979@lafn.org> wrote: >=20 >> Any ideas? >=20 > Seems you ruled out a lot of things. I'm wondering if it could be PSU = or EMI? >=20 > Also, is the asymetry in errors based on number of errors, or = percentage of errors of total packets in the direction? >=20 > Could the difference be explained by one interface receicing more, and = the other sending more, and the error percentage thus being about the = same? >=20 >> Unfortunately these servers are a 3 hour drive away at 1 AM. It=E2=80=99= s considerably longer during the day as I have to drive through Los = Angeles. >=20 > Next time you're there, probably doesn't hurt to check the wiring. Do = both run close to anything magnetic, such as PSUs, wall warts etc? >=20 > Oh, and in situations like these, I try to work in a habit of taking = pictures. Nice to have as a reference, without having to visit.=20 The input on one NIC has more packets in error and fewer total packets. = The output has more total packets and fewer error packets. The same = holds for the other NIC. There are several orders of magnitude in the = percentages. The cables are something I didn=E2=80=99t think about. I = will have them changed and see if that affects anything. =20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?28F04ECC-AF7C-4089-808B-51807CD110A9>