Date: Mon, 7 Oct 2019 17:56:07 +0200 From: =?UTF-8?Q?Morgan_Wesstr=c3=b6m?= <freebsd-database@pp.dyndns.biz> To: freebsd-questions@freebsd.org Subject: Re: Intermittent connectivity loss with em(4) Message-ID: <eaa4fb4e-308d-ae37-4ede-7901034e6644@pp.dyndns.biz> In-Reply-To: <094A1786-98A8-4CEF-A036-5809D9705EC8@mail.sermon-archive.info> References: <ba24db9b-6a34-76d4-7c95-4c2e64d0a5a0@pp.dyndns.biz> <94B563F6-55C4-46BC-BD79-5CC2AD86E6C1@mail.sermon-archive.info> <43f0decc-d459-4ded-dc46-c249bc8e24c9@pp.dyndns.biz> <094A1786-98A8-4CEF-A036-5809D9705EC8@mail.sermon-archive.info>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2019-10-07 05:55, Doug Hardie wrote: > Forgot to include, I use a lot of bge devices and some rl. However, I just discovered one of the remote client's computers has 2 em nics. I am running the ping test now on it. I did find the following in messages: > > zgrep em /var/log/messages* > /var/log/messages.6.gz:Sep 29 15:22:33 brain kernel: em0: link state changed to DOWN > /var/log/messages.6.gz:Sep 29 17:08:14 brain kernel: em0: link state changed to UP > /var/log/messages.6.gz:Sep 29 17:09:46 brain kernel: em0: link state changed to DOWN > /var/log/messages.6.gz:Sep 29 17:09:49 brain kernel: em0: link state changed to UP > > The first outage was a complete outage to that site. I don't know what caused it or how it got resolved. I never noticed the second one. > > The client has not been really thrilled with this machine. He had a much better one, but it was too noisy and replaced it. > > Yet another idea: get mtr "my trace route". package mtr-nox11 and use it to go to google. It is traceroute and ping combined. You can use -i 60 with it also. > > -- Doug Thank you for your suggestions, Doug. My suspicions right now points to my cable provider. I have two identical machines. One of them is always hooked up as my router/firewall and while that one is in service I'm free to experiment on the other. Earlier this year I switched from an ADSL provider to a DOCSIS 3.0 Cable provider. The ADSL provider used to give me 5 external IPs and I never had any trouble running both machines simultaneously through the ADSL modem. My cable provider gives me 4 IPs but this is the first time I attempt to run both machines simultaneously through the cable modem. Tcpdump shows a peculiar behaviour which is why my suspicions now points to the cable provider. While connectivity is established on my test router, tcpdump properly shows all packets leaving and entering that machine. But as soon as connectivity is lost, tcpdump suddenly starts showing packets that are leaving the _other_ machine instead. It only shows the leaving packets though, not the incoming ones but regardless I believe this behaviour isn't normal. As soon as connectivity returns on my test machine, tcpdump once again returns to only show the packets from that machine. Connectivity is never affected on the main machine when this happens and at some point I'll have to unplug it and only have the test machine connected but I have quite a few services running on that machine so it's painful to disconnect it, even for a short time. My DOCSIS knowledge is very limited and I have no idea how I can test for proper functionality of my cable modem. There are very few settings in it, none that I find relevant to this problem and the logs don't show any errors so I'm lost for now. /Morgan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?eaa4fb4e-308d-ae37-4ede-7901034e6644>