Date: Tue, 20 Apr 2021 09:39:21 -0700 From: Freddie Cash <fjwcash@gmail.com> To: Mark Delany <n7w@delta.emu.st> Cc: FreeBSD Hackers <freebsd-hackers@freebsd.org> Subject: Re: Various problems with 13.0 amd64 on vultr.com Message-ID: <CAOjFWZ4X3w1Yys5k%2BBbfnYDqm29%2B5r-Gtbt46vCvB-42suEAbQ@mail.gmail.com> In-Reply-To: <0.2.0-final-1618896757.688-0xb6a34e@qmda.emu.st> References: <0.2.0-final-1618742820.474-0x878fa2@qmda.emu.st> <20210420021318.GB18217@blisses.org> <0.2.0-final-1618896757.688-0xb6a34e@qmda.emu.st>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Apr 20, 2021 at 12:34 AM Mark Delany <n7w@delta.emu.st> wrote: > On 19Apr21, Mason Loring Bliss allegedly wrote: > > > I haven't seen a hang yet, but the test system hasn't been up much more > > than ten minutes, so I'll report back later. > > I think I've isolated it to natd traffic. > > The system stays up reliably with natd disabled but hangs within a couple > of minutes of an > inbound ipv4 traffic. > > If I just run with the ipfw rule and the divert kernel module, then no > problem the system > runs albeit without any real ipv4 traffic working for obvious reasons. But > I can happily > do anything I like in ipv6 and it runs fine. > > But as soon as natd is run with inbound traffic such as an ssh session, > then the system > mostly hangs and according to the vultr console, it's spinning at 100% CPU. > If you re-write your rules to use the in-kernel libalias support instead of divert sockets sending traffic to natd, does it stay up while passing IPv4 traffic? That would help narrow it down even further to natd issues. -- Freddie Cash fjwcash@gmail.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAOjFWZ4X3w1Yys5k%2BBbfnYDqm29%2B5r-Gtbt46vCvB-42suEAbQ>