[Gluster-devel] How to make out-of-sync files visible in replication setup

[Stephan von Krawczynski]

Hello all,

during the latest discussion I always wondered how an admin of some
replication setup can find out easily if the glusterfs is in-sync or at least
mostly synced or which files are really out-of-sync without using getfattr on
_all_ files.
Since you can have millions of files it is obvious that ls -lR over the whole
tree has chances to never return because of the pure run time.
So it would be great to have some method of finding only the out-of-sync files
inside the glusterfs.
And exactly for that case I would like to make this proposal:
How about creating an option to specify a directory on each server, where the
server itself (i.e. glusterfsd) creates a softlink to every file he sets the
xattrs to a value that implies some of his fellow servers is out-of-sync.
As soon as this xattr gets cleared the softlink is removed.
Of course the softlink should point to the full path and not a relative one.
This would allow to get a replication in-sync simply by stating the files
pointed to by the softlinks.
I guess that this idea is easy to implement and trivial in usage. I hope my
explanation is good enough, if not feel free to ask anything.
What do you think?

-- 
Regards,
Stephan