Re: fsync() fails under NFS, right?

New Message Reply About this list Date view Thread view Subject view Author view Attachment view

From: braver@pobox.com
Date: 09/14/99-03:31:29 PM Z


From: braver@pobox.com
Date: Tue, 14 Sep 1999 16:31:29 -0400
Message-Id: <199909142031.QAA03954@roll.setup.org>
Subject: Re: fsync() fails under NFS, right?


Dave Noveck states --
  o It sees the WRITE requests
  o that the server should generate (and possibly a COMMIT).
  o A server would have to try very hard to make fsync not 
  o work.  If it is not doing WRITE and COMMIT's properly,
  o lots of stuff wouldn't work.  It is easier for a 
  o client to have an fsync bug but if you think you 
  o have one, you need to demonstrate it with a 
  o small test program and a packet trace showing that
  o the correct WRITE requests are not in fact being 
  o issued.

It occurred to me this is the meat of the
argument.  You say it's all client business.  OK,
my client wants to do fsync().  Now your server
receives WRITE requests.  HOW does my client know,
whether your server WROTE THROUGH to disk or
CACHED my block?  How can I be sure that if your
server crashes, the thing written before an
fsync() doesn't die asleep in that remote cache?

It seems to us our NFS servers have huge caches
and silently swallow our small 100 MB file.  :-)
(All our machines, Linux or Sun, start at 0.5 GB
RAM.)

-- 
Cheers,
Alexy V. Khrabrov -- www.suffix.com -- braver@pobox.com


New Message Reply About this list Date view Thread view Subject view Author view Attachment view

This archive was generated by hypermail 2.1.2 : 03/04/05-01:47:34 AM Z CST