[PATCH] Allow partition size to be define relative to RAM size

[Michael Tautschnig]

Hi Julien,

I'll reply to this one first; one general remark applies to your other patches
as well: Would you mind adding those patches to the experimental branch
yourself? I don't know whether you already have commit rights in the FAI
repository, but I guess Thomas could take care of this pretty much
immediately. Obviously I'd still take a curious look at all the patches, but I'd
really appreciate if it would become a lot more visible that it's by far not me
alone anymore who works on improving setup-storage.

Now for this specific patch:

> Hi,
> 
> The attached patch (code + doc) allows partition sizes in setup-storage
> to be defined relative to the RAM size.

That's pretty cool: Another user already requested that in #493237, which got
tagged wontfix as I wasn't truly interested in that myself which caused a lack
of motivation to come up with proper support.

> 
> For some of the setups I use, I need to be able to define the size of
> one or more swap partitions as a function of the amount of RAM installed
> in the machine.
> 
> I figured it would be better to integrate this into setup-storage rather
> than playing ugly games with the disk config files and FAI hooks.
> 
> Config example:
> 
> disk_config /dev/sda fstabkey:uuid
> primary    swap    RAM:150%            swap    -
> primary    swap    RAM:50%-RAM:100%    swap    -
> primary    swap    -RAM:100%           swap    -
> logical    swap    10G-RAM:100%        swap    -
> logical    swap    RAM:100%-20G        swap    -
> 

IMHO this looks really good.

> When using mixed ranges, be careful wrt the min < max
> constraint. Currently there is no specific handling for this case, not
> sure there is a need for it.
> 

If min < max you'll get an "internal error" of setup-storage as the make_range
function, which will eventually process the values, does check for this. Not
truly nice, but at least we should not end up with some completely broken
system.

[...]

From my POV the patch looks really good; I'd prefer if you could add it to the
experimental branch yourself, but if you don't like to I'll take care of that.

Best,
Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
Url : http://lists.uni-koeln.de/pipermail/linux-fai-devel/attachments/20100526/b647a8a7/attachment.bin