Daniel Gryniewicz

This depends on the FSAL. All we do is pass through the attributes. Which FSAL are you using?

Or just turning it off at build time (-DUSE_FSAL_LIZARDFS=OFF) if you don't need it.

5.6 and 5.7 have fixes for memory usage in libcephfs. Please try 5.7 and see if that fixes your problem.

No, the proper solution would be to use CLOCK_MONOTONIC for cases that cannot go backwards. However, the *real* proper solution is to run NTP on all servers, and to start...

There are lots of things an administrator could do to a server that could temporarily or permanently mess with Ganesha's operation. I agree that changing the time is one of...

What would we do if there was a failure? Fail the client? What is the relative importance of recovery vs. getting actual work done?

This is presumably something set in your ceph.conf file. The Ganesha Ceph FSALs use the Ceph configuration, in addition to ganesha.conf.

The message comes from the ceph libraries. The only ganesha config that FSAL_CEPH has that's related to auth is the secret_access_key, which is just passed into the ceph code as...

I'm not an expert on CephFS, so I can't say for sure. But my basic understanding was that the secret_access_key referenced a key from the ceph keyring, and that that...

Error 32 is EPIPE. For TCP, this means the socket was disconnected. I assume that the NFS stack is supposed to reconnect when the server restarts. @ffilz Any ideas?