[20913] in SIPB IPv6
Re: limekiller pkgsrc update
daemon@ATHENA.MIT.EDU (Quentin Smith)
Thu Dec 31 11:32:52 2009
Date: Thu, 31 Dec 2009 11:32:48 -0500 (EST)
From: Quentin Smith <quentin@MIT.EDU>
To: Greg Troxel <gdt@work.lexort.com>
cc: sipbv6@mit.edu
In-Reply-To: <smuk4w39s7i.fsf@linuxpal.mit.edu>
Before we invest too much time in updating limekiller, I'd like to see it
move to newer x86 hardware, either as a VM or dedicated. Would NetBSD
still be the best OS choice on x86?
--Quentin
On Thu, 31 Dec 2009, Greg Troxel wrote:
> I update packges on limekiller to recent head pkgsrc (which is in a
> quarterly freeze). I discussed this with danjared@ and/or raeburn@ long
> ago and finally got to it.
>
> The big change is that quagga is now at 0.99.15, which I have been
> running at BBN and on sunpal7.
>
> Other pkgsrc changes people probably don't care about too much:
>
> mk.conf adjusted to
> build binary packages for dependencies
> install rc.d scripts in /etc
> turned off use of ccache - circular dependencies in pkg_install
> turned off X11_TYPE=xorg as that is no longer supported
>
> pkg_rolling-replace installed
>
> everything but gmake and ccache has been rebuilt (new version, or
> dependency updated)
>
> arla deleted (wasn't running, had some complicated dependencies)
> of course easy to add back
>
> perl now at 5.10
> bash not at 4
> subversion deleted, and language bindings and apache removed
> subversion-base remains, with apr1
>
> random stuff that didn't look intentional pruned
>
> other sysadmin changes:
>
> etcmanage installed, unchanged files registered
>
> added tick.ir.bbn.com (truetime XL-DC, in my office at BBN, reachable
> over IPv6) to ntp.conf
>
> future plans:
>
> update along netbsd-3 branch to latest
> maybe dropping notion of MKFOO=no - doesn't seem necessary
> probably via build done elsewhere
>
> install new bootblocks from netbsd-3
>
> move to netbsd-5, maybe via netbsd-4 (this is tricky with boot block
> changes)
>
>
> Let me know if any of this is objectionable; I'm guessing it all seems
> fine.
>
>