Re: NFSv4 Advisory vs. Mandatory locking issues

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

From: Stevan Steve Allen (scallen@us.ibm.com)
Date: 01/23/03-09:24:59 AM Z


Subject: Re: NFSv4 Advisory vs. Mandatory locking issues
Message-ID: <OF53538004.4A2FD3F9-ON88256CB7.0013178B@us.ibm.com>
From: Stevan Steve Allen <scallen@us.ibm.com>
Date: Thu, 23 Jan 2003 07:24:59 -0800




Thanks David,  for what I am asking I summarized your post as follows.

For I/O requests in an NFSv4  mandatory locking environment many NFS server
implementations may serialize the I/O requests by acquiring a file mutex.
Mutex describes serialization semantics where each object successfully
completes.  To support mandatory locking solutions, the NFSv4 protocol
introduced locked based failure conditions not supported by the mutex
serialization model for which I/O requests are failed. If an I/O request
passes the locking conditions described in the NFSv4 draft a standard NFS
server implementation will satisfy the requests locking requirements.
Additional lock based failure conditions other than those described in the
NFSv4 draft to support customer solutions such as strong locking are
considered non-standard NFS implementations. Non-standard deviations of the
NFSv4 locking protocol should be documented by the implementing product.
The NFSv4 protocol does not support or restrict the use of non-standard NFS
implementations to satisfy customer solutions.

My view may be narrow in this case, I assume the external behavior for
"implicit locking" is not agreed upon.

Thanks,
Stevan C. Allen


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 : 03/04/05-01:50:47 AM Z CST