Veritas Storage Foundation and High Availability Solutions Release Notes

Veritas Storage Foundation™ and High ... - SORT - Symantec Veritas Storage Foundation™ and High ... - SORT - Symantec

d1mj3xqaoh14j0.cloudfront.net
from d1mj3xqaoh14j0.cloudfront.net More from this publisher
17.08.2015 Views

172Upgrading to 6.0.3Upgrading to 6.0.316 Check the readiness of the nodes where you plan to upgrade. From thedirectory that contains the extracted and untarred 6.0.3 rolling patch binaries,change to the directory that contains the installmr script.Start the pre-upgrade check:# ./installmr -precheck node1 node2 ... nodeNwhere node1 and node2 are nodes which are to be upgraded.See “About the installmr and the uninstallmr scripts” on page 12.Resolve any issue that the precheck finds.17 Start the upgrade.# ./installmr [-rsh] node1 node2 ... nodeNReview the output.18 Enter the following command on each node to take service groups online:# hagrp -online service_group -sys nodename19 If necessary, reinstate any missing mount points in the /etc/vfstab file oneach node.20 Import all the diskgroups that are not managed by VCS:# vxdg import diskgroup21 Restart all the volumes by entering the following command for each diskgroup that are not managed by VCS:# vxvol -g diskgroup startall22 If you stopped any RVGs in step 10, restart each RVG:# vxrvg -g diskgroup start rvg_name23 Remount all CFS/VxFS file systems on all nodes:# cfsmount mountpointor# mount -F fstype blockdevice mountpoint

Upgrading to 6.0.3Upgrading to 6.0.317324 Remount all Storage Checkpoints on all nodes:# cfsmount /checkpoint_nameor# mount -F vxfs -o ckpt=name blockdevicemountpoint25 Start all applications which are using VxFS/CFS files systems that are notmanaged by VCS. Use the application native commands to start theapplications.Performing a full upgrade to 6.0.3 on an SF Oracle RAC clusterThe following procedure describes performing a full upgrade on an SF for OracleRAC cluster.To upgrade to 6.0.3 on an SF Oracle RAC cluster1 Make sure you have downloaded the latest software required for the upgrade.See “Downloading required software to upgrade to 6.0.3 ” on page 163.2 Log in as superuser.3 Verify that /opt/VRTSvcs/bin is in your PATH so that you can execute allproduct commands.4 From any node in the cluster, make the VCS configuration writable:# haconf -makerw5 Enter the following command to freeze HA service group operations on eachnode:# hasys -freeze -persistent nodename6 Make the configuration read-only:# haconf -dump -makero7 If Oracle Clusterware is not controlled by VCS, enter the following commandon each node of the cluster to stop Oracle Clusterware:# $CRS_HOME/bin/crsctl stop crs8 Stop VCS.# hastop -all

Upgrading to 6.0.3Upgrading to 6.0.317324 Remount all <strong>Storage</strong> Checkpoints on all nodes:# cfsmount /checkpoint_nameor# mount -F vxfs -o ckpt=name blockdevicemountpoint25 Start all applications which are using VxFS/CFS files systems that are notmanaged by VCS. Use the application native comm<strong>and</strong>s to start theapplications.Performing a full upgrade to 6.0.3 on an SF Oracle RAC clusterThe following procedure describes performing a full upgrade on an SF for OracleRAC cluster.To upgrade to 6.0.3 on an SF Oracle RAC cluster1 Make sure you have downloaded the latest software required for the upgrade.See “Downloading required software to upgrade to 6.0.3 ” on page 163.2 Log in as superuser.3 Verify that /opt/VRTSvcs/bin is in your PATH so that you can execute allproduct comm<strong>and</strong>s.4 From any node in the cluster, make the VCS configuration writable:# haconf -makerw5 Enter the following comm<strong>and</strong> to freeze HA service group operations on eachnode:# hasys -freeze -persistent nodename6 Make the configuration read-only:# haconf -dump -makero7 If Oracle Clusterware is not controlled by VCS, enter the following comm<strong>and</strong>on each node of the cluster to stop Oracle Clusterware:# $CRS_HOME/bin/crsctl stop crs8 Stop VCS.# hastop -all

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!