Date: Wed, 11 Dec 2013 12:23:23 +0600 From: "Eugene M. Zheganin" <emz@norma.perm.ru> To: freebsd-net@freebsd.org Subject: ipsec: a weird guess Message-ID: <52A804DB.8060402@norma.perm.ru>
next in thread | raw e-mail | index | archive | help
Hi. I have two CARP'ed routers servicing gre/ipsec tunnels to another FreeBSD/i386, running 8.2-STABLE (since November 2011). 22 days ago I upgraded their ipsec peer to FreeBSD 10.0-BETA1/amd64. Two days ago I upgraded one of their DNS (yeah, it gets weirder) peers from 10.0-CURRENT/amd64 (March 2013) to 10.0-BETA1/amd64. And (finally) yesterday I started getting "panic: double fault" on both target routers. Like 4-5 minutes from start. One goes down, another get in, and goes down too. I know that technical mailing lists get a lots of crazy e-mails, and this sounds like another one, but in the past I was reporting at least two major ipsec issues (and they were fixed), so I'm not that crazy. These double faults are repeatable on both routers (they boot in and go down), first I thought that may be they both got a power outage, and bgfsck is causing panic, and so on - but panics stopped when I switched off ipsec and let the clear gre run. Now they are working. I upgraded one to 9.2-RELEASE-p2/i386, but this didn't help (same double fault). Is it worth reporting ? Double fault panic is similar to ones that can be found in google (only these rocket science numbers are different). I'm pretty sure it will go away after upgrading to 10.0/amd64. If someone would tell me that story I'd say "it's crazy and impossible" too. Thanks. Eugene.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?52A804DB.8060402>