Frank Filz

Looks like this is the issue discussed somewhat over a year ago: https://lists.nfs-ganesha.org/archives/list/[email protected]/thread/JSADKLKF3GKNCKPFV5NYEJ3ML3YU7UN7/

About a year and a half ago, I did some performance testing of Kerberos with iozone. I found krb5i to be consistently pretty close to AUTH_SYS in performance with no...

Bumping this so we don't lose it. We may be able to put some resources on this in the new year.

Good point, please check that.

7541e15805ff9259f5c4d6c6fdca5c866f1cd959 in V4-dev.23 has us remove unclaimed filesystems when doing populate_posix_file_systems. That should take care of the issue., I think that patch could be back ported. It might require a...

Please submit this via Gerrithub as per src/CONTRIBUTING_HOWTO.txt

Is there still interest in this patch? If so, please submit via gerrithub

I'm not sure what's going on here. I just tested using FSAL_VFS and all squash is working as expected. Does the root of the ceph filesystem have 0777 (rwxrwxrwx) permissions?

Sure, we could implement a count. Would you like to submit a patch? Maybe we should also implement a short delay so we don't need to retry so often?

I thought about doing this, but I don't recall the parameters of why we did the retry. I'm not sure how long we have to retry.