lkml.org 
[lkml]   [1998]   [Dec]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: nfsroot pathology
Brian Craft wrote:
>
> I'm stumped on an odd nfsroot behavior I'm seeing:
>
> NFS moves at about 8 kbaud. A 1k packet every second or so. It takes
> hours to boot (it's never made it all the way up). This is on 100bT
> media.
>
> This doesn't happen if I boot the client off the hard drive & nfs mount the
> same directories. In that case it runs at a normal speed.
>
> What difference would there be in the networking between nfsroot mounting
> and mounting after the machine is up? On an isolated hub?
>
> Even stranger, on two occasions -- without changing anything -- after running
> slow for about a minute, it suddenly started running at normal speed (and
> went through the boot sequence very quickly until it choked on something
> else I haven't sorted out yet). This sudden change of speed by many orders
> of magnitude seemed completely random.
>
> The client in question is a Versa 486 with a pcmcia Linksys Etherfast 10/100.
> I'm booting from floppy with an initrd that holds the pcmcia & network
> modules.
>
> The server is a 586 with a 3com905.
>
> I've tried this with kernels 2.0.35 and 2.0.36 (on both ends).
>
> Anyone know what could cause this? I've been messing with the NFS parameters,
> but haven't settings which affect this. It appears that the first major
> timeout occurs before anything hits the wire: tcpdump on the server shows
> no activity (not even an ARP) before the 1st "NFS server not responding"
> message (with retrans=3).
This really sounds like an IRQ issue. Maybe your pcmcia setup is
different. Compare /proc/interrupts & /proc/ioports with both the normal
(off hd) and nfsroot. Are there any differences? Put a shell on your
initrd, and see if nfs mounting your nfs server works ok. What about
other networking? (ping, ftp, telnet etc). Can you ping your laptop
after it gets assigned an ip during a nfsroot boot?

Chances are that this is just a misconfiguration.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:46    [W:0.039 / U:0.196 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site