Re: Issues with using fanotify for a filesystem indexer

[Jan Kara <jack@suse.cz>]

On Fri 03-04-09 08:44:03, Alexander Larsson wrote:
> On Thu, 2009-04-02 at 21:52 +0200, Jan Kara wrote:
> > On Thu 02-04-09 19:15:37, Alexander Larsson wrote:
> > > Ah, I see. The indexer sets the flag. 
> > > I see some issues. First of all, writing the flag/mtime to disk seems
> > > like a bad idea. It'll cause a lot of writing when the indexer recurses
> > > throught the filesystem, similar to atimes. But, if you're not
> >
> >   There's some cost but it's not nearly as bad as with atimes. 
> True, its not as bad as atimes. But it still does some writes, and
> writes seem to affect i/o performance more than low prio reads from the
> indexer. I'm very wary about the background indexer process disturbing
> the foreground processes. This is one of the main problems with current
> indexers.
  I agree - that's why I have beagle turned off on my system ;)

> > > persisting the flag/mtime then you need to keep all the dentries with
> > > the flag set in memory, which has resource use risks similar to
> > > unbounded event queues.
> >   Ah, true - I have implemented just the persistent case and have not
> > thought too much about the non-persistent one. You're right that it won't
> > work because we'd pin memory.
> 
> So, where do you persist the flag/time? Is there some availible space in
> the inode for it on ext3/4?
  Yes, there's enough space in ext3/ext4 inode. I've already talked about
it with Ted and other fs developers at Plumbers Conf. and they weren't
opposed to such change.

									Honza
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR