(687.1 8-4-1998)) id 41256739.00611321 ; Fri, 19 Mar 1999 18:40:18 +0100
Date: Fri, 19 Mar 1999 18:44:49 +0100
From: georg.tasman_at_db.com
Subject: Follow up: nfs-Problem
To: tru64-unix-managers_at_ornl.gov
Message-id: <41256739.0061120E.00_at_dbogw2-e1.esb.eur.deuba.com>
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii
Content-disposition: inline
X-Lotus-FromDomain: ZENTRALE_at_DEUBAINT
Hi Managers,
the only answer I got was from Nick Hill. He experienced the same problems as
ours.
His errors disappeared when he introduced patchkit 3.
We did the same, but our errors still remain.
the number of events increase the longer the systems run.
Does anybody else has the same or a similar problems :
Writing to a nfs-filesystem yields "Interupted system call" or I/O-Error
The only sign of the error in the system is an badcall in nfsstat
all standard errorfiles are clean.
Thank you for your replys
Georg Tasman
Original question as follows:
Hi Managers,
at the moment we got a strange nfs-problem:
the clients mount some fs from the nfs-server (all 4.0D with patchkit 2)
sometimes the clients can not write to existing files. The write stops at
different locations with an
write error. Neither the error-logfiles on the client nor on the server have
entrys.
nfsstat shows some Client nfs badcalls.
extract of nfsstat:
Client rpc:
tcp: calls badxids badverfs timeouts newcreds
0 0 0 0 0
creates connects badconns inputs avails interrupts
0 0 0 0 0 0
udp: calls badxids badverfs timeouts newcreds retrans
85079 6 0 68 0 0
badcalls timers waits
74 25 0
Client nfs:
calls retrans badcalls nclget nclsleep
85011 60 6 85009 0
Client nfs V2:
null getattr setattr root lookup readlink read
0 0% 1 0% 0 0% 0 0% 0 0% 0 0% 0 0%
wrcache write create remove rename link symlink
0 0% 0 0% 0 0% 0 0% 0 0% 0 0% 0 0%
mkdir rmdir readdir statfs
0 0% 0 0% 0 0% 292 0%
Client nfs V3:
null getattr setattr lookup access readlink read
0 0% 984 1% 13 0% 2479 2% 323 0% 0 0% 13931 16%
write create mkdir symlink mknod remove rmdir
44317 52% 52 0% 0 0% 0 0% 0 0% 353 0% 0 0%
rename link readdir readdir+ fsstat fsinfo pathconf
3 0% 0 0% 35 0% 40 0% 4460 5% 3 0% 0 0%
commit
17725 20%
does anyone has a clue, how to prevent these badcall ?
do i have to increase nfsiod on the client or nfsd on the server ?
or do i have to change some kernal-parameters ?
or do i have to use different mount options ? (actual options:
hard,rw,retry=5,bg)
Thank you
Georg Tasman
Received on Fri Mar 19 1999 - 17:43:30 NZDT