home *** CD-ROM | disk | FTP | other *** search
- X-Gateway-Source-Info: INTERNET
- Path: sparky!uunet!elroy.jpl.nasa.gov!usc!news.service.uci.edu!unogate!mvb.saic.com!tgv.com!info-multinet
- Date: 20 NOV 92 21:05:38 GMT
- Newsgroups: vmsnet.networks.tcp-ip.multinet
- X-Return-path: <info-multinet-relay@TGV.COM>
- X-RFC822-From: adelman (Kenneth Adelman) @ TGV.COM
- From: adelman@TGV.COM
- Subject: Re: NFSMOUNT in a cluster
- Organization: The INFO-MULTINET Community
- Message-ID: <36A00FB320NOV92210538@TGV.COM>
- Nntp-Posting-Host: Mvb.Saic.Com
- Lines: 20
-
- > >> What is possible and will we run in trouble with shared writes in case the
- > >> clusterworde mounting does not exist ?
- > >
- >> Record locking doesn't work across NFS, but file locking will if
- >> the NFS Server supports a lock manager.
-
- > I don't think the original question was in reference to shared writes
- > between VMS and UNIX. I think it was coordination between members of a
- > cluster writing to NFS client files that they all can see. Since the
- > cluster members don't see the same device (apparently you cannot mount
- > NFS devices cluster-wide?) you probably don't get cluster-wide distributed
- > lock-manager locking to coordinate shared writes amongst cluster members,
- > right? In other words, it doesn't act like a physical disk with multiple
- > paths or a single-path disk device served to the cluster and mounted
- > cluster-wide.
-
- Unfortunately, you're not wrong.
-
- Ken
-