On 2013-09-25, Darren Spruell <[email protected]> wrote: > On Tue, Sep 24, 2013 at 5:18 AM, Stuart Henderson <[email protected]> > wrote: >> On 2013-09-24, Darren Spruell <[email protected]> wrote: >>> On Fri, Sep 20, 2013 at 12:13 PM, Alexey E. Suslikov >>><[email protected]> wrote: >>>>> Any idea what the issue could be? >>>> >>>> could you provide netstat -s output after "several hours"? >>> >>> To circle back, errors started building again and below is netstat -s. >>> As Chris suggested I'll try a snapshot and see if issue surfaces >>> again. >> >> Are you certain your cabling and switch are OK? > > No. I've got to swap them out to see but wanted to first check out the > firewall due to fact that rebooting it clears the issue for a few > days. > >> What do you mean by "ethernet tap"? > > One of these units: > > http://auctionimages.s3.amazonaws.com/1936/20572/14577108.jpg > > It sends a copy of network traffic from link between firewall and LAN > switch to an IDS sensor. I'll also need to verify it along with the > switch (it's a layer 1 device and passes through link negotiation but > could still introduce errors I believe). > > > +-------------+ +----------+ +-----+ +--------+ >| cable modem |----| firewall |----| tap |----| switch | > +-------------+ +----------+ +-----+ +--------+ > | | > | | > | | +--------+ > | \.| | > \___| sensor | > | | > +--------+ > >
I thought so, but then I noticed you were using autoneg and I was under the (possibly mistaken) impression that these didn't work properly with autoneg. (though I'm also a bit unsure whether sis works properly *without* autoneg as I seem to remember some problem in that area ;)
