[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-arm-kernel
Subject:    Re: NFS: server not responding still trying with D-Link DES-1008D
From:       "Prakash Kamliya" <prakash.kamliya () matrixtelesol ! com>
Date:       2006-02-28 9:51:07
Message-ID: 002a01c63c4a$d2727240$5f32a8c0 () Matrix
[Download RAW message or body]

Dear Pavel,
Thanks for your reply. I tried what you suggested

>
> Hello Prakash,
>
> please, can you test and report results from next two experiments
>
> 1) add "tcp" option to the "nfs" mount options
>    (this is our temporary solution of the problem)

I tried with TCP option for NFS, but it didn't work in both cases when
problem with switch and when with problem with switch, it failed giving
these messages.

Root-NFS: mountd port is 1026
NFS: nfs_mount(c0a80146:/home/prakash/.")
RPC: error 5 connecting to server 192.168.1.70
VFS: unable to mount root FS via NFS, trying floppy.

> 2) test "ping -f -s 20000 server_ip" from the client
>    (this length seems to have allmost 100% packet lost)
>

I also tried with ping, with NFS on UDP, as on TCP it didn't worked in
normal condition also. Setup is like this Embedded PBX Ethernet port
(192.168.1.225) on which I am porting Linux is connected to D-Link DES-1008D
switch, other port of this switch is connected to my Destkop PC running
Win-2000 (192.168.1.71) and third port is connected to my local R&D LAN
network on which I am running Linux server running on Red-Hat 9.0
(192.168.1.70).

When in normal condition I mean PBX-Kernel is able to mount NFS
successfully,
ping (with -f and -s 2000 option) from Linux server to PBX is 100% lost, and
ping from Linux server to Desktop PC 0% lost.

But when in abnormal condition, I mean when my PBX-Kernel is not able to
mount NFS successfully and gives error NFS: still trying. Even pinging from
my Linux server to Destkop PC, packet lost is 100% ! .


If I remove all of these three interfaces from switch and reinsert only my
desktop PC and LAN port (not inserting my PBX Ethernet port), in that case
also packet lost between my Destkop PC and Linux server is 100 %

> We are experiencing exactly similar problems as you describe
> on x86 based university lab setup. Server is Fedora-2,
> clients are Debian. Problem experienced with most combinations
> of Redhat-2.6.8 kernel and Vanilla 2.6.15.1/4 kernels.
>
> We have not these problems before and we need to find
> if this is network infrastructure related
> or latest kernel version related problem.
>
> Thanks
>
>                 Pavel Pisa

Funny thing is if I power off and on D-link switch, it works fine. 0% packet
loss !
I personally feel if it is a problem with Linux server or Destkop PC, it
should not start working normal if I do restart of D-link switch.
>From this observation I personally feel that switch gets into some state in
which packet lost is very much more!, and the only way to come out from this
state for switch is restarting it. I may be wrong also!

Thanks
- Prakash


-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php
[prev in list] [next in list] [prev in thread] [next in thread]