Major issues with nfsv4

[J David]

Using oneopenown in even limited testing seems to cause NFS mounts to
completely hang (requiring "umount -N") on a fairly regular basis.

The problems without oneopenown are worse than that it will
"eventually constipate the mount."  There also seems to be some sort
of garbage collection or something that occurs periodically that
causes one nfscl process to jump to 100% CPU and create systemwide
disruption, causing delay in many unrelated system calls (30+ second
delays observed in at least fork, systemctl, and ioctl via ktrace of
processes that do not touch any NFS filesystem).  As far as I can
tell, stuff like procstat -ka doesn't run until it's already over, at
which point there's nothing to see.

Sorry I don't have more information about that.  Unfortunately, we
discussed the status of this today, and my update-to-NFSv4 project has
basically been killed; we're backing off to NFSv3.

Thanks for the help and information and all the hard work on NFS for
many many years!