Oscar sets up exporting for /home on the masternode automatically. It sounds like you have not exported the /home1 and /home2 drives. Make sure that they are listed in the /etc/exports file on their originating computer. Look at the /etc/exports file on the master node for an example. You will also have to fiddle with pfilter to let the traffic through. I would personally just off pfilter (assuming you have other security available or are not connected to the internet directly), but that is mainly because I have never bothered to find out how it works in detail.
Just a suggestion, I would not name the other directories /home1 and /home2, as they probably not be used for user data in the same way as /home is. I would name them something like /scratch1 and /scratch2 or /nodedrive1 and /nodedrive2 or something. This is a somewhat unusual setup, from my limited experience, which is why I did not understand what you were trying to do initially. Having three active NFS servers could generate a lot of traffic on your network if nothing else. Then if you get up to 16+ nodes it would be a mess... There are some clustered file systems floating around out there if you want to use all the disks in a fairly transparent way and minimize network traffic. It really depends a lot on what you are trying to do, and how many nodes you plan to have in the long run. On 10/13/06, Ivan Ivanov <[EMAIL PROTECTED]> wrote: > Hi Michael, > > It might be my fault not explaining too well. Here is a different > description of the problem: > > ======================================================= > The cluster has 3 nodes (n0 - master; n1 and n2 - slaves) and each has > its > own hard drive, named: /home, /home1 and /home2. We were able to mount > /home > to n1 and n2 so the cluster runs ok (we use MPI). > > The problem is that we could not access /home1 and /home2 from n0, when > trying to use the hard drive for storage. Here is the situation: > when in n0: read /home YES; read /home1 NO; read /home2 NO > when in n1: read /home YES; read /home1 YES; read /home2 NO > when in n2: read /home YES; read /home1 NO; read /home2 YES > > The 'mount: RPC: Program not registered' error msg pops out when we > tried to > mount /home1 or /home2 to n0. > ======================================================= > > Hope this helps, > > Ivan > > > > >>> "Michael Edwards" <[EMAIL PROTECTED]> 10/10/06 4:31 PM >>> > The mapping of home drives is from the master node to the compute > nodes, so it sounds like your set up is working as I would expect it > to. > > NFS takes care of returning changes made on the nodes back to the > server automatically, you don't need to mount the home directories of > the compute nodes back to the server, the data should simply stay in > sync as long as the clocks are all synchronized. > > Perhaps I am not understanding your problem completely though. > > On 10/10/06, Ivan Ivanov <[EMAIL PROTECTED]> wrote: > > We are running Oscar on a 3 node (16 processor) DELL cluster with > > RedHat > > 2.4.18-4smp. > > We have the following problem: > > > > When attempting to map the home directories of the 2nd and 3rd nodes > > to > > the first one (the master) the mount fails, and we get the message > > 'mount: RPC = Program not registered' > > On the other hand side, mounting the home directory of the 1st node to > > the > > second and the third nodes is fine. > > > > Any suggestions are welcome. > > > > Thanks, > > > > Ivan Ivanov > > > > > > > > > > > ------------------------------------------------------------------------- > > Take Surveys. Earn Cash. Influence the Future of IT > > Join SourceForge.net's Techsay panel and you'll get the chance to > share your > > opinions on IT & business topics through brief surveys -- and earn > cash > > > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > > _______________________________________________ > > Oscar-users mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/oscar-users > > > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share > your > opinions on IT & business topics through brief surveys -- and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > Oscar-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/oscar-users > > > ------------------------------------------------------------------------- > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job easier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 > _______________________________________________ > Oscar-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/oscar-users > ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Oscar-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oscar-users
