[Gluster-devel] inodes

[Emmanuel Dreyfus]

Amar Tumballi <amarts at gmail.com> wrote:

> For 3.2.x release branch the applied patch is not enough. Grep for similar
> patch in master branch (grep for ia_ino in git log's output). That would
> show you that the culprit here is itransform in cluster translators.
> There are patches in queue to get in for fixing this issue in release
> 3.2 branch at review.gluster.com

I am alost done: afr does afr_itransform on inode numbers from readdir,
but not those from lookup/stat. What is the correct behavior? 

Also, is it desired that you treat READDIR and READDIRP differently?
Speaking of READDIRP, it should be prefered over READDIR for the native
client, shouldn't it?

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu at netbsd.org