Re: BTRFS in 2012?

On Jan 22, 2012, at 11:27 AM, Maarten Vanraes wrote:

>> If BTRFS is used by default on boxes, then file system snapshots
>> insturmented in RPM's (or URPMI's) state machines become possible.
>> 
>> (aside)
>> There's a closely related usage case in build systems based on
>> populating a chroot which then uses a depsolver to populate
>> the chroot based on BuildRequirements. In the build system
>> case, reverting to the original chroot snapshot is (perhaps:
>> a measurement is needed to say for sure) an efficient implementation
>> to return to the original image (yes I am making an assumption that
>> each chroot is also a mount point.)
>> 
>> Is BTRFS being considered in 2012? If not, then RPM/URPM
>> using snapshots is less important than otherwise.
>> 
>> 73 de Jeff
>> 
> 
> for mageia, me and tmb are looking at btrfs, but it's just not good enough
> yet imho...

Thank you for the info (and consistent with Fedorable features where
BTRFS keeps getting pushed to the N+1 release for several years now).

Note that BTRFS mounts for build system chroot's might be feasible
even if BTRS is still not ready for general deployment.

(changing topic while I am here, related to file systems but not BTRFS)

There is another file system called hugetlbfs that has usage cases
in the "cloud" because the page size increases dramatically (to >4Mb
per page) so that multiple VM's can share large amounts of data easily.

Is there interest in hugetlfs in 2012?

More directly
	How much specific interest is there in "cloud" functionality in 2012?

Mandriva has traditionally been a "desktop linux" with a KDE speciality
niche and a lot of attention to "polish" detail to increase usability.

As the "cloud" deploys, then the feature set for Mandriva is going to change,
adding support for "hugetlbfs" is just one of many implementations that
will become increasingly important.

Is there interest in "hugetlbfs" in 2012?

Perfectly okay to say *shrug* these are obscure questions without obvious answers.

hth

73 de Jeff