Frank Filz

Submitted.

This issue has had no activity for several months, and it seems like it's an issue external to Ganesha development. Closing as unresolved.

Let's see some thread backtraces for the threads in syscall. Is it possibly a mutex syscall and then maybe a deadlock?

We had a problem with the code config, see https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1173557?usp=search

On the stalled connection, the client must not be submitting new requests (you could confirm this with tcp trace). It looks like your sfs code is waiting on a condition...

Do you have urcu installed properly? [root@fedora src]# rpm -qa | grep userspace-rcu userspace-rcu-0.12.1-3.fc34.x86_64 userspace-rcu-devel-0.12.1-3.fc34.x86_64

Hmm, I'm not familiar with Ubuntu, it looks like there's some issue with the urcu headers, maybe a different version that Ganesha is expecting?

You might want to look into using the async capabilities. If you need them for more than read and write we could expand the set. It IS OK for a...

Is this still an issue?

We probably do need to add some documentation for ntirpc... Understanding that no one ever has time to document code design... The thing to look for is are you somehow...