Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 29 Oct 2008 10:27:48 -0700
From:      Chris St Denis <chris@smartt.com>
To:        freebsd-jail@freebsd.org
Cc:        stable@freebsd.org
Subject:   Re: 7.x and multiple IPs in jails
Message-ID:  <49089D14.7040603@smartt.com>
In-Reply-To: <20081029072821.S2978@maildrop.int.zabbadoz.net>
References:  <Pine.OSX.4.64.0810280227350.4630@toasty.nat.fasttrackmonkey.com> <487086DA-4514-44E7-AB9F-F1D98C652980@yellowspace.net> <49078377.2090807@smartt.com> <20081029072821.S2978@maildrop.int.zabbadoz.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Bjoern A. Zeeb wrote:
> On Tue, 28 Oct 2008, Chris St Denis wrote:
>
> Hi,
>
> [ jail patches ]
>
>> Serious question here (not trolling).
>>
>> These patches have been around for years, why have they never been 
>> committed to trunk/stable?
>
> Well, the multi-ipv4 patch has been for a while - what we are talking
> about at the moment is more.
>
> If you look at older status reports they said soemthing like "there is
> the need for this at the moment but it's not considered to be the
> right thing".
>
> There are multiple reasons for that, that I can think of:
>
> 1) some larger parts (of the network stack|kernel) get plastered with
>    all kinds of if (this) if (that) checks complicating code, making
>    it unreadbale, having to be maintained, not ignored for security, ...
>    It's important to really catch all the places, .. which it seems we
>    had been doing well though not 100% well as I just found out
>    currerntly preparing more if (this) if (that) checks for something
>    not really important but still being a problem - since the first
>    day it turns out.
>
> 2) there is questionable logic in them and while we had been living
>    with it up to now, it came up during review process for the commit
>    to HEAD (so it could be merged to stable) and it turns out that
>    properly solving it isn't a easy or simple task and multiple people
>    have been pondering over this for days now. Even after removing
>    some optional code paths for simplicity things are still not always
>    definite in what would happen.
>
> 3) <just insert more other stuff here>
>
>
> Nonetheless they are very helpful and very usable (else I wouldn't
> have worked on it).
>
> The plan as the status report will say is to get this in, merge it to
> stable/7 before 7.2  and keep it in 8.
>
> 8 will also have vimages and ideally I'd like to see this entire jail
> IP hacks be gone for 9, when vimage will provide the infrastructure,
> etc.  This means that 8 would be the transition period. But that's
> just me and my ideas - we'll see how it'll go.
>
>
> /bz
>
Thanks for the info from all who responded.

I hadn't heard of vimage before, but after doing some searching on it it 
sounds like it will be very good improvement to jails.

If we can get resource limits on jails too in a near future release, 
Jails will become a competitive solution for VPS systems.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?49089D14.7040603>