Re: Réf. : Re: alas for thee, EVMS

What type of locking do LVM2 utils use?  Does in-kernel device-mapper
provide locking?

In pondering what EVMS3 architecture would look like, it occurrs to me
that the mismatch between LVM2 and EVMS's use of device-mapper is
currently a major impediment to EVMS maintenance... being more like LVM2
(or LVM2 being more like EVMS) would give more users/maintainers.

If you look at EVMS from DOS partitions and above, it can emulate LVM2
(compatibility volumes) , or it can add a layer on top (evms metadata
for volume names and other features).  So, with different handling of
device nodes, things would interoperate fairly well.

Going below the DOS partitions, LVM2 really has nothing, relying
completely on the in-kernel partition to disk mapping code.  EVMS OTOH,
doesn't need it, and just uses raw disks + device-mapper.  So, to move
things forward, it seems that it is in fact LVM2 that needs enhancement,
specifically to be able to at least distinguish device-mapper mappings
that are DOS partitions, and consume and otherwise utilize them 100% the
same as the in-kernel /dev/sda1 /dev/sda2 block devices.

In the style of LVM2 + md raid + encfs etc, there would be a separate
utility to discover and setup the dos partition mappings, for LVM2 to
use.  It would go into initrds, same as md raid setup etc.  In fact, it
could even be evms_activate, with an evms.conf that prohibited discovery
past the dos partition legacy devices, and didn't just put the device
nodes in /dev/evms/.

I'm sure a stripped down EVMS or a tool based on it's DOS partition
discovery code, could be made easily enough, the same tool could
integrate with udev as well.; the focus then becomes LVM2 being usable
without in-kernel partitioning.

Cheers,

Jeremy

On Mon, 2008-12-01 at 09:34 +0100, [email protected] wrote:
> It would be also great to remove the main lock in /var/lock/evms-engine and
> to consider a fine-grained locking system. We've had numerous cases where
> the only possible solution to use evms was to reboot the whole server
> before.
> 
> I think it's the weakest part of EVMS.

-- 
Jeremy Jackson
Coplanar Networks
(519)489-4903
http://www.coplanar.net
[email protected]


-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Evms-devel mailing list
[email protected]
To subscribe/unsubscribe, please visit:
https://lists.sourceforge.net/lists/listinfo/evms-devel