Re: partman-auto-lvm patches
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Joey Hess wrote:
> Fabio Massimo Di Nitto wrote:
>
>>in a previous message to this list from Colin Watson
>>(http://lists.debian.org/debian-boot/2005/09/msg00209.html) a set of patches
>>were made available to make partman-auto-lvm (pal from now) more complete
>>and somehow working.
>>
>>Unfortunatly i did delete the sequencial patches by mistakes. I am sorry about
>>it but i have prepared again a full diff with explanations of why stuff has
>>been implemented in a certain way.
>
>
> Where is the wayback machine or google cache when you need them. Oh
> well, thanks for doing the work again.
eh I think Colin managed to recover some of them, but they were not all there.
>
> Some questions:
>
> - Is the partman-auto-lvm setup modifyable with lvmcfg afterwards?
once pal complete the works, you are pushed back to the manual partition menu to
confirm the changes. IF i didn't made any mistake it should work just fine.
I see no reason why lvmcfg shouldn't handle it.
> - IIRC powerpc is not the only arch where parted doesn't understand
> lvm. Should we disable it for the others too?
Yes. I only had a limited set of arches to test on.
> - One of the things I hd wondered about using LVM is if it would be
> better to not use up all the free space in the VG by default, and
> leave some fraction for when the user decides to grow one of their
> partitions. Any thoughts/experience on that.
IME i see no point in leaving unused free space. If you need more space, you
plug extra storage. Either ways this is one of these grays areas where users
will never be happy. My *cough*authoritative*cough* decision was to take the
whole space.
> - Are you gonna help us maintain this going forward? :-)
I see no big problems with it, but be aware that i am not a d-i expert or have
much spare time to follow debian-boot ;)
>
>>Please note that there is a lot of polish that can/must be done. When I started
>>hacking on pal, i didn't realize how much code duplication was pulled in and
>>definetely there are several bits that can be designed better.
>
>
> FWIW, the reason for most of that duplication is that partman-auto-lvm
> was written in the middle of the sarge release process and we didn't
> want to make the obvious changes to partman-auto and risk breaking it,
> so duplicated the code instead. It would be good to get rid of that
> duplication.
>
Yes i agree, but the other components were a bit too scary for me to touch close
to our release. I was hoping to get some help by the relative maintainers.
>
>>The easiest would be to add a flag that will make that recipe looks like:
>>
>>8 1 1 boot-prep
>> $primary{ }
>> $bootable{ }
>> method{ prep } .
>> $nolvm{ }.
>>
>>(flag is just an idea and name up for discussion... so please don't take this as
>>absolute statement, as well it can be the other way around.. canbelvm{ } flag on
>>the other partitions ;))
>
>
> This seems like a good approach to me.
>
Ok. So I think you should decide what the flag should be named and to what
partition apply them to.
>
>>As last note about the patch automatically_partition/vg_all_free/choices has
>>been disabled because of http://bugzilla.ubuntu.com/show_bug.cgi?id=15017
>>but perhaps there might be work arounds for it that i am not aware of.
>>Mostlikely one you d-i guru's will have an answer.
>
>
> Not offhand.. if the it's not working I'd favour removing it entirely
> until fixed and adding a TODO item.
>
Agreed.
Fabio
- --
I'm going to make him an offer he can't refuse.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org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=bC+d
-----END PGP SIGNATURE-----
Reply to: