We don't use networkmanager but configure the interfaces through /etc/network/interfaces all the time.. But we still have this problem.
I figured out, that autofs-mounts have the correct clientaddr set while fstab mounts don't. Our fstab-mounts also have clientaddr=0.0.0.0. But when I manually unmount them and remount them, they get the right clientaddr but still the problem exists, at least we still have the error-messages in dmesg shown.
We don't use networkmanager but configure the interfaces through /etc/network/ interfaces all the time.. But we still have this problem.
I figured out, that autofs-mounts have the correct clientaddr set while fstab mounts don't. Our fstab-mounts also have clientaddr=0.0.0.0. But when I manually unmount them and remount them, they get the right clientaddr but still the problem exists, at least we still have the error-messages in dmesg shown.
I also found this redhat bugreport https:/ /bugzilla. redhat. com/show_ bug.cgi? id=732748 which claims that the problem is fixed with these patches: http:// article. gmane.org/ gmane.linux. nfs/48705 .
I used kernel 3.3.6-030306- generic now but we still have the "nfs4_reclaim_ open_state: Lock reclaim failed!" messages in dmesg but no hang so far..
Does your workaround still work for you? Do you had any hangs or the error-messages in dmesg anymore?
Does anyone know what the clientaddr-field is used for?