Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 29 Jan 2001 13:36:33 +0100 (CET)
From:      Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
To:        <freebsd-questions@freebsd.org>
Cc:        <admin@dbai.tuwien.ac.at>
Subject:   NFSv3 problems
Message-ID:  <Pine.BSF.4.31.0101291319350.56113-100000@taygeta.dbai.tuwien.ac.at>

next in thread | raw e-mail | index | archive | help
We recently switched a couple of 4.2-RELEASE boxes from NFSv2/UDP to
NFSv3/TCP (which is the default now).

NFS server is a Solaris 2.6 box with sufficient memory and horse power
for our use.

However, whenever the server has heavy IO (CPU states: 20.8% kernel, 79.2%
iowait) due to backups etc., we are seeing errors like the following when
accessing the NFS server (for a compile job):

  nfs server vexpert:/files3: not responding
  nfs server vexpert:/files3: is alive again  (after fractions of a second)
  nfs server vexpert:/files3: not responding
  plan.o: Interrupted system call
  /var/tmp/cck28604.s: Assembler messages:
  /var/tmp/cck28604.s:205327: FATAL: Can't write plan.o: Interrupted system call

or

  ... disk fill (when in fact, it is not).

These failures are strictly confined to FreeBSD clients with NFSv3 (either
TCP or UDP). FreeBSD clients with NFSv2/UDP do not experience this, nor do
Solaris clients.

So, this seems to be a FreeBSD bug. What shall we do?

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.31.0101291319350.56113-100000>