[CentOS] NFS problem: getattr ERROR: Request couldn't be completed in time

Tue May 31 17:15:51 UTC 2016
Matthias Leopold <matthias at leopold.priv.at>

i posted this also to linux-nfs, maybe this is a more appropriate place 
as it's clearly a "user" problem

###

hi,

i'm trying to mount a NFS 4 share with "sec=sys" from a NetApp FAS3250 
on a CentOS KVM guest. this doesn't work, the mount doesn't complete and 
hangs.

mount command is:
/bin/mount -v -t nfs -s -o vers=4.1,sec=sys 10.110.96.90:/home /home/cir

what i'm seeing in tcpdump is:

08:36:48.875019 IP 10.110.96.112.2456702824 > 10.110.96.90.2049: 212 
getattr fh 0,1/53
08:36:48.875355 IP 10.110.96.90.2049 > 10.110.96.112.2456702824: reply 
ok 96 getattr ERROR: Request couldn't be completed in time

server: 10.110.96.90
client: 10.110.96.112

status of nfs services:

nfslock: running
rpcbind: running
rpcidmapd: stopped

this happens with a centos 6 _and_ a centos 7 client (both KVM guests).
NFS 3 mount of the share doesn't work either.

other kvm guests (centos 6) on the same physical host can mount the 
share with no problem.

the netapp admin tells me there's "no problem" on his side.
iptables is disabled of course.

###

what could be the problem on my (client) side?
could this be a kvm problem?

kvm host is centos 6.
network mode is "bridge".
network driver is "virtio".

thx for any advice
matthias