SLOW NFS-READ: linux2.2 client IRIX server

SLOW NFS-READ: linux2.2 client IRIX server

Post by Thomas Butsche » Thu, 30 Dec 1999 04:00:00

Hi folks,

I've a very strange problem with the nfs client of the 2.2 kernel.

A file mounted from an IRIX 6.* NFS server takes much too long time to
be copied by linux 2.2.* to the local disk. xosview shows 40K, a 1.8M
file takes 45 secs to be copied! I tried every option with mount (noac,
rsize, wsize...) but nothing changed this effect. I tried this with a
10Mbit net, but the same occurs with 100Mbit.

The same situation on linux 2.0.* is fine (xosview shows 1.1M, even with
no additional rsize option).

What the hell changed from 2.0 to 2.2 (last kernel I checked was 2.2.13)
that could cause this strange effect? What's even move strange is that
if I create some net traffic between these two hosts (e.q. 'll' the
mounted dir) then the data bursts a normal speed, as long as this
'noise' lasts.

Can anyone give me a hint how to solve this problem? I even would invest
some time in looking at the kernel code but I don't have any idea where
to start in fs/nfs. I compared fs/nfs of 2.0 with that of 2.2 but really
a lot seems to have changed so I can't see THIS difference in the code.

I'm the adminstrator of a big linux-cluster and changing the kernels
down to 2.0.* is not my favorite.

Thanks for any hint.


1. HELP: Slow NFS client to server causing Very slow NFS install

My RH 5.2 NFS Install is SO SLOWWW!!

This is taking forever!!  The link speed currently is
284 MB in 17 hours!! or 4.6kbytes/s ~ 46kbps

Why is there such a slow server to client link?

I'm doing this using Lynksys Etherfast 10/100 card on a PII350 desktop
a P75 laptop with the PCMCIA version of the Etherfast 10/100.  The
LEDs on the cards both indicate that the connections is at 100Mbps, but

See below for all of my network config files in /etc.

I believe that with host.conf setup for "order hosts,bind" that I don't
need a
nameserver.  I don't have a local name server setup.  

I have tried modifying some of the parameters below during the
running exportfs and turning eth0 off and on (using usernet on the
DURING THE INSTALLATION (it seems that RH5.2 mounts "hard" during NFS
ie it is continuing on after I turn eth0 back on) but with no change on

I did notice when I set the network up before the install that pinging
client to server was odd, some packets took 10ms, others took 1000ms in
almost every other packet trade off.  The ping from server to client was
the order of 10 ms every packet.  Now with the install going on, ping
server (desktop) is averaging 350ms with any packet size from 256Bytes

I have the following setups in exports, host.conf, hosts, hosts.deny,
hosts.allow and resolv.conf:

 cat exports
/mnt/cdrom laptop(ro)  
/home/chris laptop(rw)

cat host.conf
order hosts,bind
multi on

 cat hosts   desktop   laptop       localhost       loopback

 cat hosts.deny
portmap: ALL

 cat hosts.allow

 cat resolv.conf

I have tried resolv.conf with and without commenting out the nameservers
local isp nameservers,) but with no effect.

2. Adaptec 2940 - Supported?

3. Linux NFS client and IRIX NFS server

4. HELP: Critical Errors from Harddisk

5. IRIX NFS server and Linux NFS client

6. memmove problems

7. NFS woes: tcsh's internal ls fails on Solaris 2.6 client / IRIX 6.x server

8. Network Card Driver for Linux

9. nfs: Irix client, DEC server

10. NFS weirdness, IRIX 6.3 server to Solaris 2.6 client

11. Need info on PC/NFS and Unix NFS servers, server/client ratio

12. nfs tuning for fast server, slow client

13. Slow NFS server -> high client load av