Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 04 Apr 2012 08:00:46 -0700
From:      perryh@pluto.rain.com
To:        jameschen@juniper.net, fbsd8@a1poweruser.com
Cc:        questions@freebsd.org
Subject:   Re: Questions about Jail
Message-ID:  <4f7c621e.tnQwV40ucBoiDtJs%perryh@pluto.rain.com>
In-Reply-To: <4F7B19CE.80401@a1poweruser.com>
References:  <079BB83C1486C245B64B055CF016336A02612843@emailhk3.jnpr.net> <4F7B19CE.80401@a1poweruser.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Fbsd8 <fbsd8@a1poweruser.com> wrote:

> In most cases your jail environment will function ok as long as
> its the same base release level. Example, host=8.0 jail1=8.1 and
> jail2=8.2

IIUC, a better example would be host=8.2, jail1=8.1 and jail2=8.0.
A point release is not supposed to make any incompatible changes to
the kernel ABI, but it might add new interfaces not present in the
older kernel.

> But host=8.2 and jail1=9.0 will have unknown reliability.

I would say it is only an accident if (jail major > kernel major)
works, because the KABI will likely have changed between N.x and
(N+1).x.  However, host=9.0, jail1=8.x should work if the host
kernel includes the COMPAT_FREEBSD8 option.

> Technically there is no checks stopping someone from doing this
> and from the outside all will look correct, but it will fail and
> you may lose both the host and jail.

You may indeed lose the jail, but if _anything_ done in the jail is
able to corrupt the host there is by definition a bug in the host's
jail support.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4f7c621e.tnQwV40ucBoiDtJs%perryh>