Frank Filz

V2.8.2 isn't even the latest version of the V2.8 stream. Any problem in Ganesha would require a new version to pick up the fix (unless you back ported the fix...

How easy is this to recreate? We could crank up the debug, but if it takes a while to recreate, it could be an enormous debug log. Instead of the...

Are you following the guidelines here: https://doc.beegfs.io/latest/advanced_topics/nfs_export.html All of that will apply to Ganesha also (well no worry about the file lease bit, Ganesha will close files when the NFSv4...

If the problem is a handle that is too big for Ganesha to use with FSAL_VFS (which is any handle too big for NFSv3) then Ganesha will be giving the...

Ganesha doesn't guarantee it can always work the same as Kernel NFS. When FSAL_VFS was written, no Kernel filesystem known to us used such a large handle so we just...

> > Ganesha doesn't guarantee it can always work the same as Kernel NFS. When FSAL_VFS was written, no Kernel filesystem known to us used such a large handle so...

What version of Ganesha are you using? Also, are you using dbus to add an export to a running Ganesha, or configuring nested exports in your config file? What FSAL...

Ah, OK, so you are trying to put c1 and c2 under c in the pseudofs but they are beside c in the GPFS file system? In order to do...

I don't think that's right. get_gsh_export_by_pseudo_locked is looking for an export by pseudo path. Who is calling get_gsh_export_by_pseudo_locked? Maybe that's the wrong search to call in this path? I'll try...

No it needs to be get_gsh_export_by_pseudo there because we are building the pseudofs. Hmm, could you also try this one with 4.0 or next? This is definitely doing something tricky...