Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!sdd.hp.com!caen!destroyer!news.itd.umich.edu!news.itd.umich.edu!not-for-mail
From: pauls@terminator.rs.itd.umich.edu (Paul Southworth)
Newsgroups: comp.os.386bsd.questions
Subject: NFS Client Problems
Date: 5 Jul 1993 21:31:27 -0400
Organization: Information Technology Division, University of Michigan
Lines: 41
Distribution: world
Message-ID: <21aklf$eir@terminator.rs.itd.umich.edu>
NNTP-Posting-Host: terminator.rs.itd.umich.edu
This is on a NetBSD, 486SX, 8Mb system, with one user on. I generally don't
use NFS server connections much, but here I wanted to mount a server volume
from my SS1, tar 40Mb of stuff onto it, then unmount it, in lieu of a tape
drive, and since I have lots of space on the Sun. After about twenty minutes,
the machine began to really crawl, pings taking up to 3000ms from the Sun (these
machines being separated by about 6 feet of wire on the same subnet, on a
holiday to boot). I killed the tar, killed the nfsiod's (there were four) and
then restarted eight new nfsiod processes and started over. Obviously the
nfsiod's are really hammering the CPU (486sx-33) bigtime, and it just can't
take it. Four vs. eight nfsiod's apparently don't matter much. It took at
least an hour for the second four of them to even wake up, and it took over
three hours for it to tar about 25Mb of the job. Other clients on the net are
perfectly happy with this server (both sun3, sun4 and NeXT systems).
Using the stock NetBSD 0.8 release, with an NE2000 on thinnet.
"top" statistics, for those interested...
load averages: 2.81, 2.77, 2.53 22:47:16
22 processes: 20 sleeping, 2 running
Cpu states: 0.3% user, 0.0% nice, 98.4% system, 1.3% idle
Memory: Free: 2112K Active: 1948K Inactive: 1404K Wired: 1488K
kill
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
118 root 2 0 80K 44K sleep 43:46 102.49% 102.49% nfsiod
119 root -1 0 80K 28K sleep 30:00 46.29% 46.29% nfsiod
159 root 65 0 292K 216K run 0:07 30.67% 23.49% top
120 root -1 0 80K 28K sleep 17:38 18.80% 18.80% nfsiod
136 root 67 0 280K 204K run 15:40 17.72% 17.72% tar
121 root -1 0 80K 28K sleep 4:07 16.99% 16.99% nfsiod
125 root 10 0 80K 28K sleep 3:01 5.57% 5.57% nfsiod
124 root 10 0 80K 28K sleep 3:14 5.27% 5.27% nfsiod
123 root 10 0 80K 28K sleep 3:28 4.98% 4.98% nfsiod
122 root 10 0 80K 28K sleep 3:49 4.74% 4.74% nfsiod
51 root 18 0 72K 32K sleep 0:22 1.07% 1.07% update
151 root 18 0 244K 192K sleep 0:00 0.00% 0.00% csh
53 root 18 0 188K 148K sleep 0:43 0.00% 0.00% crond
1 root 10 0 160K 96K sleep 0:00 0.00% 0.00% init