nfs lockd broken
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux (Ubuntu) |
New
|
Undecided
|
Unassigned |
Bug Description
I've tried AMD64 2.6.28-2.3-generic with Jaunty userspace, and with Intrepid userspace, and NFS is b0rked with both (nfs client and nfs server). Mounts work with -o nolock, though. (-o nolock doesn't require NFS lockd to be running). Otherwise, NFS mount attempts time out.
nfs-common's /etc/init.d script fails, and it runs that in postinst, so this should be reproducible for anyone, even if they don't normally use NFS. And nfs-kernel-server won't start either, even with an empty /etc/exports. (I tried running /etc/init.
This was a problem with the first Jaunty 2.6.28-1-ub-generic kernel, too. Intrepid's 2.6.27-10 is ok.
As I said, I was able to reproduce this by booting to my Jaunty chroot, where I didn't previously have nfs-common installed at all, so it's not a stale config file causing problems.
I searched around some, expecting other people would have hit this, but maybe not many people use NFS on the systems where they play around with new kernels. Otherwise, maybe I just missed finding an already reported bug, or worse, this isn't as reproducible as I'm hoping.
I can confirm this. I upgraded to Jaunty for testing last night on this Kubuntu box and noticed this bug just a little while ago.
NFS working on this machine when it was running Intrepid (mounting a share on another machine also running intrepid), but since the upgrade I can only mount using the -o nolock option like Peter.
I also installed nfs-kernel-server, and saw the same failure described by Peter.