NFSv3:
Server : AIX 5.3
Client Linux 2.6.9
Test : connectathon 2004
* All tests passed, 2 warning.
NFSv3
Server AIX 5.3
Client Linux 2.6.10
* Big file test (test5 from BASIC tests) is blocked
* Lock tests blocked (Tests 14)
For the test5 (read/write a 30Mo file) a sequence of packet is sent many times
in an infinite loop :
client to server
NFS V3 WRITE CALL, FH: 0x38ad8e6 Offset:32768 Len:38768 UNSTABLE
client to server Fragmented IP protocol (UDP, 0x11, offset 1480 )
client to server Fragmented IP protocol (UDP, 0x11, offset... )
client to server Fragmented IP protocol (UDP, 0x11, offset... )
....
client to server Fragmented IP protocol (UDP, 0x11, offset... )
client to server Fragmented IP protocol (UDP, 0x11, offset 25160)
(there is a delay of 1,4 seconds before
previous sequence is repeted)
__________________________________________________________
dmesg is telling :
nfs: server aix_server not responding, still trying
...
Since all is right with kernel 2.6.9, that looks like a regression...
Even if that is the true behaviour it may exist a way to stop this process
cleanly : since NFS umount is can't be done the client does not reboot, and I
need to turn the power off...
NFSv4
_____
Works a better...
All tests passed. Big file tests is _really_slow (many minutes to copy a 30Mo
file over a 100Mo/s network)
Vincent
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
ty den 08.02.2005 Klokka 10:22 (+0100) skreiv Vincent Roqueta:
> Since all is right with kernel 2.6.9, that looks like a regression...
> Even if that is the true behaviour it may exist a way to stop this process
> cleanly : since NFS umount is can't be done the client does not reboot, and I
> need to turn the power off...
There were no NFS or RPC client changes made between 2.6.9 and 2.6.10,
so it looks like you're going to have to look at networking changes.
Since you are seeing problems with both UDP and TCP, then I would start
by checking whether or not any changes have been made to the network
card driver.
Cheers,
Trond
--
Trond Myklebust <[email protected]>
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Le mardi 8 F=E9vrier 2005 15:49, Trond Myklebust a =E9crit=A0:
> ty den 08.02.2005 Klokka 10:22 (+0100) skreiv Vincent Roqueta:
> > Since all is right with kernel 2.6.9, that looks like a regression...
> > Even if that is the true behaviour it may exist a way to stop this
> > process cleanly : since NFS umount is can't be done the client does not
> > reboot, and I need to turn the power off...
>
> There were no NFS or RPC client changes made between 2.6.9 and 2.6.10,
Well... Some littles changes :-) =20
I tried various kernel and patches versions since this morning:
linux-2.6.9 -> PASSED
linux-2.6.10-rc1-CITI_NFS4_ALL-2.dif -> PASSED
linux-2.6.10-rc1-CITI_NFS4_ALL-3.dif -> (answer soon)
linux-2.6.10-rc1-CITI_NFS4_ALL-4.dif -> FAILED
The problem is between in the NFS4_ALL-4.dif or NFS4_ALL-3.dif ...
Vincent
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
Le mardi 8 F=E9vrier 2005 16:43, Vincent Roqueta a =E9crit=A0:
> Le mardi 8 F=E9vrier 2005 15:49, Trond Myklebust a =E9crit=A0:
> > ty den 08.02.2005 Klokka 10:22 (+0100) skreiv Vincent Roqueta:
> > > Since all is right with kernel 2.6.9, that looks like a regression...
> > > Even if that is the true behaviour it may exist a way to stop this
> > > process cleanly : since NFS umount is can't be done the client does n=
ot
> > > reboot, and I need to turn the power off...
> >
> > There were no NFS or RPC client changes made between 2.6.9 and 2.6.10,
>
> Well... Some littles changes :-)
>
> I tried various kernel and patches versions since this morning:
> linux-2.6.9 -> PASSED
> linux-2.6.10-rc1-CITI_NFS4_ALL-2.dif -> PASSED
> linux-2.6.10-rc1-CITI_NFS4_ALL-3.dif -> (answer soon)
> linux-2.6.10-rc1-CITI_NFS4_ALL-4.dif -> FAILED
>
> The problem is between in the NFS4_ALL-4.dif or NFS4_ALL-3.dif ...
>
>
> Vincent
>
Problem is in 2.6.10-rc2-CITI_NFS4_ALL-1 patch set
however linux-2.6.10-rc1-CITI_NFS4_ALL-4.dif patch set FAIL _some_ times.
I am investigating more to know what patch is wrong.
Vincent
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs