[BIG RFC] Filesystem-based checkpoint

Dave Hansen dave at linux.vnet.ibm.com
Thu Oct 30 09:36:25 PDT 2008


On Thu, 2008-10-30 at 12:25 -0400, Oren Laadan wrote:
> Dave Hansen wrote:
> > On Tue, 2008-10-28 at 15:56 -0500, Serge E. Hallyn wrote:
> >> If you like I can take a shot at whipping up the new mini-fs, though
> >> I think you're having fun :)
> > 
> > There are a couple of concepts that just get easier once you start
> > thinking of this as an entire fs too.  For instance, cr_ctx just becomes
> > crfs_sb.  For things like dumping in parallel, we get locking and
> > lifetime rules for free from the vfs.
> 
> Well, 'cr_ctx' is per-checkpoint, while crfs_sb will single for the
> entire system. So you'll need to add something per checkpoint anyway.

I was thinking of it more along the lines of requiring a new filesystem
mount for each checkpoint.  That way, we dispose of the checkpoint by
the act of unmounting.

> What other concepts get easier ?

The amount of infrastructure needed to do lookups for shared objects
goes to zero.  We don't need a hash table or ids with which we index
into that table.  Filesystems are good at giving things names and
finding them later.

-- Dave



More information about the Containers mailing list