Hi, you have to make sure that on all nodes exporting /gpfs/home that filesystem must have identical inode-structures an an identicat major/minor number for the device where the data actual resides.
You can determine the device by using 'mount | grep "/gpfs/home"'. Kind regards, Nils Hildebrand > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Geoff Dunk > Sent: Tuesday, September 09, 2008 10:29 AM > To: General Linux-HA mailing list > Subject: RE: [Linux-HA] NFS locks > > Nils > > I use the following command which resides within /etc/fstab > > HPC:/gpfs/home /gpfs/home nfs defaults 0 0 > > I do get exactly that message about stale NFS mounts. > > I am not sure quite what you mean about major and minor > numbers of devices? > > Regards > > Geoff Dunk > HighPerformance Computing Engineer > > Phone 01604 799834 > Mobile 07825 344332 > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of > Hildebrand, Nils, 232 > Sent: 09 September 2008 08:53 > To: General Linux-HA mailing list > Subject: RE: [Linux-HA] NFS locks > > Hi, > > what mount command do you use on your clients? > > Do you get error-meesages about stale NFS mounts on your clients? > > Are major and minor number of the device where your > NFS-filesystem resides the same on all your cluster-nodes? > > > Kind regards, > > Nils > > > -----Original Message----- > > From: [EMAIL PROTECTED] > > [mailto:[EMAIL PROTECTED] On Behalf Of Geoff Dunk > > Sent: Thursday, July 17, 2008 10:48 PM > > To: [email protected] > > Subject: [Linux-HA] NFS locks > > > > Can someone please help? > > > > I have 2 servers providing access to the same data and need to > > implement Linux-HA to provide failover between the 2. The > trouble is, > > when the 1st server is switched off, the NFS client doesn't > > automatically remount the second server. > > > > When accessing the HA IP when server 1 is down, I get routed to the > > 2nd , so HA is working, but NFS will not automatic switch without > > forcing a umount -a of the NFS file system on the client or > rebooting > > the client machine. > > > > Regards > > > > Geoff Dunk > > High Performance Computing Engineer > > Mercedes-Benz HighPerformanceEngines Ltd > > > > > > Mercedes-Benz HighPerformanceEngines Ltd. > > Board of Directors: Thomas Weber (Chairman), Norbert Haug, Leopold > > Mikulic, Sten Ola Kaellenius, Christian Andreas Struwe. > > Domicile and Court of Registry: Brixworth, United Kingdom. > > Registered in England and Wales - Company No. 1760288; VAT No. 387 > > 6578 82. > > Delivery address: Morgan Drive, Brixworth, Northamptonshire, > > NN6 9GZ, United Kingdom, Tel +44 (0)1604 880100, Fax: +44 > > (0)1604 882800. > > If you are not the addressee, please inform us immediately that you > > have received this e-mail by mistake, and delete it. > > We thank you for your support. > > > > _______________________________________________ > > Linux-HA mailing list > > [email protected] > > http://lists.linux-ha.org/mailman/listinfo/linux-ha > > See also: http://linux-ha.org/ReportingProblems > > > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
