Hi Ben,
Thank your for trying to help. Indeed, I can get as much information
as you might need (I am the administrator of those machines). After
enabling rpcdebug -m nfs -s mount on the client this is what I get:
[ 461.238568] NFS: nfs mount
opts='hard,sec=sys,vers=4.1,addr=10.0.2.9,clientaddr=10.1.0.12'
[ 461.243621] NFS: parsing nfs mount option 'hard'
[ 461.246573] NFS: parsing nfs mount option 'sec=sys'
[ 461.249809] NFS: parsing sec=sys option
[ 461.252364] NFS: parsing nfs mount option 'vers=4.1'
[ 461.255472] NFS: parsing nfs mount option 'addr=10.0.2.9'
[ 461.258864] NFS: parsing nfs mount option
'clientaddr=10.1.0.12'
[ 461.262610] NFS: MNTPATH: '/home'
[ 461.264757] --> nfs4_try_mount()
[ 461.273063] <-- nfs4_try_mount() = -1 [error]
when running tcpdump on the nfs server (tcpdump -i eth0 host 10.1.0.12
and port nfs -n -s 0 -vvv), and mounting on the client, this is what I
get:
tcpdump: listening on eth0, link-type EN10MB (Ethernet), capture
size 262144 bytes
09:55:01.603947 IP (tos 0x0, ttl 64, id 15242, offset 0, flags [DF],
proto TCP (6), length 188)
10.1.0.12.epp > 10.0.2.9.nfs: Flags [P.], cksum 0x1fdf
(correct), seq 1347340908:1347341056, ack 1722815553, win 1919, length
148: NFS request xid 3760933574 144 getattr fh 0,1/53
09:55:01.604325 IP (tos 0x0, ttl 64, id 38336, offset 0, flags [DF],
proto TCP (6), length 132)
10.0.2.9.nfs > 10.1.0.12.epp: Flags [P.], cksum 0x168c
(incorrect -> 0x08f4), seq 1:93, ack 148, win 1432, length 92: NFS reply
xid 3760933574 reply ok 88 getattr ERROR: unk 10016
09:55:01.604717 IP (tos 0x0, ttl 64, id 15243, offset 0, flags [DF],
proto TCP (6), length 40)
10.1.0.12.epp > 10.0.2.9.nfs: Flags [.], cksum 0xf0e6 (correct),
seq 148, ack 93, win 1919, length 0
09:55:01.612334 IP (tos 0x0, ttl 64, id 15244, offset 0, flags [DF],
proto TCP (6), length 252)
10.1.0.12.epp > 10.0.2.9.nfs: Flags [P.], cksum 0xe9ed
(correct), seq 148:360, ack 93, win 1919, length 212: NFS request xid
3777710790 208 getattr fh 0,1/53
09:55:01.612495 IP (tos 0x0, ttl 64, id 38337, offset 0, flags [DF],
proto TCP (6), length 240)
10.0.2.9.nfs > 10.1.0.12.epp: Flags [P.], cksum 0x16f8
(incorrect -> 0x417f), seq 93:293, ack 360, win 1432, length 200: NFS
reply xid 3777710790 reply ok 196 getattr NON 4 ids 0/47982174 sz
-1769090185
09:55:01.652270 IP (tos 0x0, ttl 64, id 15245, offset 0, flags [DF],
proto TCP (6), length 40)
10.1.0.12.epp > 10.0.2.9.nfs: Flags [.], cksum 0xef34 (correct),
seq 360, ack 293, win 1941, length 0
I have noticed there is this error 10016, that in nfs4.h translates to
NFS4ERR_WRONGSEC, suggesting that the server does not allow 'sys', but I
have checked again my /etc/exports file and contains
/home 10.0.0.0/8(rw,no_subtree_check,sec=sys:krb5:krb5i:krb5p)
and also is reported by the operating system:
# cat /proc/fs/nfsd/exports
/export/home
10.0.0.0/8(rw,root_squash,sync,wdelay,no_subtree_check,uuid=0743ce63:00c1185a:00000000:00000000,sec=1:390003:390004:390005)
Any ideas?
Thank you!
Felix
---
Felix Rubio
"Don't believe what you're told. Double check."
On 2020-01-22 19:30, Benjamin Coddington wrote:
On 22 Jan 2020, at 4:22, Felix Rubio wrote:
Hi everybody,
I have a kerberized NFSv4 server that is exporting a mountpoint:
/home 10.0.0.0/8(rw,no_subtree_check,sec=krb5:krb5i:krb5p)
if I mount that export with this command on the client, it works as
expected:
/sbin/mount.nfs4 NFS.domain:/home /network/home -o
_netdev,noatime,hard,sec=krb5
However, if I modify the export to be
/home 10.0.0.0/8(rw,no_subtree_check,sec=sys:krb5:krb5i:krb5p)
and I mount that export with sec=sys, as
/sbin/mount.nfs4 NFS.domain:/home /network/home -o
_netdev,noatime,hard,sec=sys
I get the following error:
mount.nfs4: timeout set for Fri Jan 17 14:11:32 2020
mount.nfs4: trying text-based options
'hard,sec=sys,vers=4.1,addr=10.2.2.9,clientaddr=10.2.0.12'
mount.nfs4: mount(2): Operation not permitted
mount.nfs4: Operation not permitted
What might be the reason for this behavior?
Hi Felix,
I don't know. Can you get more information? Try again after `rpcdebug
-m
nfs -s mount`. That will turn up debugging for messages labeled for
mount,
and the output will be in the kernel log. There are other facilities
there,
see rpcdebug(8).
Another good option is getting a network capture of the mount attempt
and
trying to figure out if the server is returning an error, or the client
is
generating the error.
There are also a lot of "nfs", "nfs4", and "rpc" tracepoints you can
enable
to get more information.
Ben
