From: Spencer Shepler (shepler@eng.sun.com)
Date: 05/23/01-06:09:37 PM Z
Date: Wed, 23 May 2001 18:09:37 -0500 From: Spencer Shepler <shepler@eng.sun.com> Subject: Re: todo list for RFC3010 Message-ID: <20010523180937.J100350@dhcp-aus08-229.central.sun.com> On Tue, Matthew J. Zito wrote: > At 1:29 PM -0500 5/22/01, Spencer Shepler wrote: > >You will find pointers to the "todo" list and a version of RFC3010 > >that is the base document for our updates at: > > > >http://www.nfsv4.org/rfc3010updates/index.html > > "- If the server exporting a filesystem read-only, should > OPEN(ACCESS_WRITE) fail with NFS4ERR_ROFS? Or will the OPEN succeed > and any WRITE()s will fail with NFS4ERR_ROFS." > > I would suggest that open should fail on a read-only filesystem. > That seems to be the way the unixes I've looked at handle read-only > filesystems. Are there reasons for having it fail on write()? Failing on open() will be a change in behavior from the way NFSv2/v3 implementations deal with this. Without OPEN, v2/v3 have no way to determine if the file is read-only. So failing the OPEN() will be a change in behavior. Is it a reasonable change? I think so but as has been pointed out in other threads these types of 'reasonable' changes start to add up such that NFSv4 will behave enough differently from v2/v3 as to "bother" the end user. -- - Spencer -
This archive was generated by hypermail 2.1.2 : 03/04/05-01:48:47 AM Z CST