Installation procedure for SC2.1 HANFS.
1. Install all recommended O/S patches to all nodes. 2. Install the the Admin Work Station cluster pakages from cdrom. #./scinstall option client (to install client pakages) 3. SETUP your environment for PATH and MANPATH on the Admin Work Station. PATH=/opt/SUNWcluster/bin: MANPATH=/opt/SUNWcluster/man: 4. Add the CRN and terminal server to the /etc/hosts on all nodes. 5. set /etc/nsswitch.conf on all nodes to files. 6. Edit /etc/serialports with clusternode networkterminalserver portnumber eg phys-ukc1 nts 5002 7. Edit /etc/clusters with clustername nodename nodename eg ukc phys-ukc1 phys-ukc2 8. Test on the AWS using #ccp clustername & 9. Using the cluster console setup PATH and MANPATH for both nodes; PATH=/opt/SUNWcluster/bin:/opt/vxva/bin:/opt/SUNWpnm/bin MANPATH=/opt/SUNWcluster/man:/opt/SUNWvxva/man:/opt/SUNWpnm/man DISPLAY=AWS:0.0 TERM=whatever 10. Check the storage array cabling and power to arrays and setup OBP scsi-initiator-id on one or both nodes for any nodes using multipacks. This prevents initiator conflict when using multi-initiator. 11. Check controller number are the same on both nodes when ssaadm,luxadm or format is run on both nodes 12. Ensure that you know how the private network interfaces will be named in Solaris as you will need this info for the scinstall on the CRN's 13. Install the cluster s/w to both CRN at the same time using the common window scinstall > server > install > automatic This will set in your CDB the number of nodes in the cluster. The info to configure your private networks, the flag to use a ccdvol and your quroum device. It will also install VM packages. 14. Select the data services you require from the list during the install. 15. Patch the CRN and the AW with the 2.1 cluster patch (105458-09) 16. Patch the CRN with the VM2.5 patch (105463-05) 17. Reboot all nodes If SCI is used the CRN must be power cycled to load f/w onto SCI cards 18. If using SCI then the s/w has to setup. Use one of the template file from /opt/SUNWsma/bin/Examples that reflects your config or edit the file to reflect your config. Run the /opt/SUNWsma/bin/smconfig -f from one node only Power cycle the nodes 19. Setup VM by creating a simple disk to go into rootdg or encapsulate your boot disk when you run vxinstall. Note that vxinstall dose not need to be run if the simple disk method is used. 20. Ensure that all drives you intend to use in your cluster have a standard disk label and that none have a VM label with slice 3 (private) and slice 4 (public) set. 21. Setup the shared ccdvol using #confccdssa If setp 20 is not done confccdssa may fail or try to import the sc_dg that may be on private regions from a previous life. Use vxdiskadm, option 8 (import diskgroup) just to list diskgroups your node can see out on the arrays. 22. License the CRN using /etc/opt/license/lit 23. Start the first node into the cluster using #scadmin startcluster nodename clustername 24. Once the cluster configuration steps have finished start the other node #scadmin startnode 25. On the AW start the Cluster monitor from ccp 26. Edit /etc/hosts on all nodes including the AW with an entry for your logicalhost 27. Create a disk group with 2 disks in, one from each controller and then deport the diskgroup 28. Create the new logical host from one node only e.g # scconf (clustername) -L nfs -n(nodelist) -g hadbms -i le0,le0,hadbms e.g # scconf (clustername) -F nfs (on both nodes to create the admin file system and the edirectory structures and files to mount the admin file system) The diskgroup should be imported and the admin file system mounted 29. Check that the logicalhost network is up using ifconfig -a 30. Switch the logical host to the other node using #haswitch othernode logicalhost 31. On the node owning the logical host create a mirrored volume of 1GB in the diskgroup 32. Create mount points for the 1GB volume #mkdir /blabla 33. setup vfstab.logicalhost for the volume created, on both nodes 34. Reconfigure the cluster so it mounts the new volume and test using haswitch. 35. Edit the dfstab.logicalhost to share the file system. Set permission to anon=0 36. Register the nfs using hareg -s -r nfs -h logicalhost hareg -y nfs 37. Reconfigure the cluster to test that volumes are mounted and shared. #haswitch -r 38. Test the HA-NFS from the AW by doing cd /net/logicalhost/mountpoint and create a file or test script.