Date: Thu, 31 Aug 2017 09:10:04 +1000 From: Jason Tubnor <jason@tubnor.net> To: "Patrick M. Hausen" <hausen@punkt.de> Cc: freebsd-stable <freebsd-stable@freebsd.org> Subject: Re: Bridged networking regression in 11.0? Message-ID: <CACLnyCK41G023dSKOjhqZeb3Fo5KgbjcJe0mJr4kgBy=qBjwVw@mail.gmail.com> In-Reply-To: <E844186F-F2A2-4B78-AAC6-1C865524312B@punkt.de> References: <F70DE809-66DD-4A02-92DB-1A21FC0B017F@punkt.de> <C1B63BB4-D36F-45E8-A5FA-B3835FC7B2B8@punkt.de> <E844186F-F2A2-4B78-AAC6-1C865524312B@punkt.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On 30 August 2017 at 20:49, Patrick M. Hausen <hausen@punkt.de> wrote: > > The only real change relates to IF teardown and we did not change anything > (stop/restart jails) when the outages occured. > > So ... the network interface driver? Probably? > > > We are using igb(4) with vlan(4), tap(4) bundled into bridge(4). The taps are entries to bhyve guests. These were built initially on FreeBSD 11.0 and have been recently uplifted to FreeBSD 11.1. We have never had any issues with these bridges, even under heavy load. We do want to netmap/vale but not until it is maintainable via freebsd-update (built into generic). So based on what you have said, a bit more of a breakdown to determine the culprit that is within the bridge might be needed, especially with VIMAGE (though our FreeNAS jails use VIMAGE and bridges without breakage also). Cheers, Jason.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CACLnyCK41G023dSKOjhqZeb3Fo5KgbjcJe0mJr4kgBy=qBjwVw>