Hi
sorry if this is a duplicate, I have been having some problem
subscribing to the ml.
I am having some problem with nfs-kernel on a debian amd64 (sid) i am
using version (1:1.2.2-1).
when I do a tcpdump on my nfs machine and do a mount -o vers=4,udp
nfs:/exports/video /exports/video/ it fails
mount -o vers=3,udp nfs:/exports/video /exports/video/
mount.nfs: mount to NFS server 'nfs:/exports/video' failed: timed out, giving up
and with tcpdump, I see the request from from the client and I see the
reply ok from the nfs machine, but it seem like the nfs packet doesn't
make it to the wire, machines on the same network doesn't see the
packet.
on the nfs server I have a bridge interface br0 => eth0 eth1 eth2 eth3 vde0
on vde0 I have attached virtualbox 3.2.4, and when I look on the vde
switch with tcpdump I don't see the reply packet.
I have done some limited testing for udp and it seems to be working,
traced dns packets
I can mount tcp, but if I user vers4 i get wierd gid uid numbers
ls -l
drwxrwsr-x 2 4294967294 4294967294 4.1k Oct 25 2009 Funnies
rpcinfo -p
program vers proto port
100000 2 tcp 111 portmapper
100024 1 udp 32765 status
100024 1 tcp 32765 status
100021 1 udp 32768 nlockmgr
100021 3 udp 32768 nlockmgr
100021 4 udp 32768 nlockmgr
100021 1 tcp 32768 nlockmgr
100021 3 tcp 32768 nlockmgr
100021 4 tcp 32768 nlockmgr
100003 2 tcp 2049 nfs
100003 3 tcp 2049 nfs
100003 4 tcp 2049 nfs
100227 2 tcp 2049
100227 3 tcp 2049
100003 2 udp 2049 nfs
100003 3 udp 2049 nfs
100003 4 udp 2049 nfs
100227 2 udp 2049
100227 3 udp 2049
100005 1 udp 32767 mountd
100005 1 tcp 32767 mountd
100005 2 udp 32767 mountd
100005 2 tcp 32767 mountd
100005 3 udp 32767 mountd
100005 3 tcp 32767 mountd
on nfs restart
Jun 13 22:05:20 max kernel: [21723.836493] svc: failed to register
lockdv1 RPC service (errno 97).
Jun 13 22:05:20 max kernel: [21723.837490] NFSD: Using
/var/lib/nfs/v4recovery as the NFSv4 state recovery directory
Jun 13 22:05:20 max kernel: [21723.837507] NFSD: starting 90-second grace period
tcpdump on nfs machine of mount -o vers=4,udp nfs:/exports/video
/exports/video/
tcpdump -pni br0 host 192.168.10.221 and not port 22
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on br0, link-type EN10MB (Ethernet), capture size 65535 bytes
22:09:12.565697 IP 192.168.10.221.30209 > 192.168.11.10.53: 52743+ A?
nfs.hme1.samad.com.au. (39)
22:09:12.565891 IP 192.168.11.10.53 > 192.168.10.221.30209: 52743*
2/2/3 CNAME max.lan1.hme1.samad.com.au., A 192.168.11.10 (182)
22:09:12.567227 IP 192.168.10.221.35835 > 192.168.11.10.53: 59898+ A?
nfs.hme1.samad.com.au. (39)
22:09:12.567369 IP 192.168.11.10.53 > 192.168.10.221.35835: 59898*
2/2/3 CNAME max.lan1.hme1.samad.com.au., A 192.168.11.10 (182)
22:09:12.568676 IP 192.168.10.221.243695162 > 192.168.11.10.2049: 40 null
22:09:12.568726 IP 192.168.11.10.2049 > 192.168.10.221.243695162:
reply ok 24 null <<<<<<<<<<<<< this is missing on the client ....
22:09:13.668091 IP 192.168.10.221.243695162 > 192.168.11.10.2049: 40 null
22:09:13.668158 IP 192.168.11.10.2049 > 192.168.10.221.243695162:
reply ok 24 null
22:09:15.867266 IP 192.168.10.221.243695162 > 192.168.11.10.2049: 40 null
22:09:15.867345 IP 192.168.11.10.2049 > 192.168.10.221.243695162:
reply ok 24 null
22:09:17.680551 IP 192.168.10.221.58336 > 192.168.11.10.25826: UDP, length 917
Help :)
On Tue, 2010-06-15 at 12:01 +1000, Alexander Samad wrote:
> Hi
>
> sorry if this is a duplicate, I have been having some problem
> subscribing to the ml.
>
>
>
> I am having some problem with nfs-kernel on a debian amd64 (sid) i am
> using version (1:1.2.2-1).
>
> when I do a tcpdump on my nfs machine and do a mount -o vers=4,udp
> nfs:/exports/video /exports/video/ it fails
UDP transport is not supported by the NFSv4 protocol. Some servers may
allow it, but most will not.
Cheers
Trond
On Tue, Jun 15, 2010 at 11:15 PM, Trond Myklebust
<[email protected]> wrote:
> On Tue, 2010-06-15 at 12:01 +1000, Alexander Samad wrote:
>> Hi
>>
>> sorry if this is a duplicate, I have been having some problem
>> subscribing to the ml.
>>
>>
>>
>> I am having some problem with nfs-kernel on a debian amd64 (sid) =A0=
i am
>> using version (1:1.2.2-1).
>>
>> when I do a tcpdump on my nfs machine and do a mount -o vers=3D4,udp
>> nfs:/exports/video /exports/video/ it fails
>
> UDP transport is not supported by the NFSv4 protocol. Some servers ma=
y
> allow it, but most will not.
okay, but why does the nfs-kernel on linux look like it is sending
packets (via udp), but they are not hitting the wire.
>
> Cheers
> =A0Trond
>
>
On Thu, Jun 17, 2010 at 10:42 PM, Alexander Samad <[email protected]> w=
rote:
> On Tue, Jun 15, 2010 at 11:15 PM, Trond Myklebust
> <[email protected]> wrote:
>> On Tue, 2010-06-15 at 12:01 +1000, Alexander Samad wrote:
>>> Hi
>>>
>>> sorry if this is a duplicate, I have been having some problem
>>> subscribing to the ml.
>>>
>>>
>>>
>>> I am having some problem with nfs-kernel on a debian amd64 (sid) =A0=
i am
>>> using version (1:1.2.2-1).
>>>
>>> when I do a tcpdump on my nfs machine and do a mount -o vers=3D4,ud=
p
>>> nfs:/exports/video /exports/video/ it fails
>>
>> UDP transport is not supported by the NFSv4 protocol. Some servers m=
ay
>> allow it, but most will not.
>
> okay, but why does the nfs-kernel on linux look like it is sending
> packets (via udp), but they are not hitting the wire.
Okay I have vers=3D4, tcp working.
But I would like to have UDP, I had issues with laptop that were
asleep (hibernating) and then unable to reconnect when I restart the
laptop - sometimes because the nfs server needed a reboot etc... With
udp I don't have these problems.
So i have done some more testing.
I have enabled nfs on another server, with the same package revisions
as my primary nfs server. Say server max is original nfs the one not
working and nas is the newer one.
if I do this on nas
exportfs -o async,ro 192.168.8.0/22:/mnt/tmp
I can do this on max
mount -o vers=3D4,udp,ro,async nas:/mnt/tmp/ /mnt/tmp/
but I can't do the reverse.
I have tried rebooting max into difference kernel revisions -
currently using 2.6.34, tried 6.32 and 6.30 - all failed.
the pain is this used to work.
What I see is the same flow of packets but from max I see the outbound
ack, but it doesn't hit the wire.
which makes me sort of think this might not be he right ml :( but I
tried different kernels....
thanks
frustrated
>
>
>>
>> Cheers
>> =A0Trond
>>
>>
>