ROOT MIRRORING IN VERITAS VOLUME MANAGER PDF

ROOT MIRRORING IN VERITAS VOLUME MANAGER PDF

If you are really interested the Blueprints book 'Boot Disk Management' is a worthwhile investment although it dows cover a lot of the same ground as the PDF file. The article also contains some suggestions for creating unencapsulated mirrored root environments to make maintenance of your Veritas disks easier as well. The Sun Blueprint link above describes this all far better than I have under the Clone Disks section and includes a script to do it all for you. This means you can stil maintain your Veritas disks, make changes to the Veritas boot disk in the event that for some reason you cannot boot from one of the Veritas mirrors. Edits made to /etc/system (remove rootvol from vxio and from rootvol) and /etc/vfstab to mount /dev/dsk devices for root disk mean it is not under Veritas control, BUT Veritas will be started. Having mirrored the disk, we detach the plexes and remove them from Veritas. Second disk will be a point in time mirror of the first - not encapsulated since there was no data on it to begin with. JB RE: alternate boot disk Annihilannic (MIS) 12 Mar 03 06:20 Have a look at using the Veritas persistent Data Change Object combined with snapshotting to enable a really quick resynchronisation. Test boot on contingency disk 'ok> boot altboot' check root is mounted as a standard not a VxVm device. 'vxedit set reserve=on altboot' - prevent VxVM from usingДЇsck all the partitions since they will be dirtyД¬reate a file in /mnt indicating when copy was takenД®dit /mnt/vfstab with new device names cNtNdNs0,1.Д®dit /mnt/etc/system, delete rootdev references to prevent VxvM trying to start on contingency disk boot 'vxplex -o rm dis rootvol-0? swapvol-0?' - remove plexes 'vxprint | grep altboot' - note subdisk and plex names Using eeprom command add altboot devalias to nvramrc and set use-nvramrc? to true '/etc/vx/bin/vxbootsetup altboot' - make it bootable add any other (/opt, /var) from your root disk as above 'vxassist -g rootg mirror swapvol altboot' - add swap VxVM must copy all of the data on the original disk over to the mirror. '/etc/vx/bin/vxrootmir altboot' - mirror root partition If the volumes are not mirrored back to the root disk in the same order they. 'vxedit set nconfigall nlog=all rootdg' - keep VxVM config on all rootdg disks 'vxdg -g rootdg adddisk altboot=cNtNdN' - add to rootdg '/etc/vx/bin/vxdisksetup -i cNtNdN' - init new disk If disk has been added in-flight without rebooting. You will need another disk preferably on a different controller as large, or larger than your existing root disk paritions. #metattach d10 d11 Step 10: Check sync status.Suggest you have a mirror disk and alternative boot disk for best resilience. #metainit d11 1 1 /dev/rdsk/c1t1d0s1 Step 9:Add new metadevice to mirror. **If you didnвЂ™t delete metadevices at step 2, you donвЂ™t need this step. #metadevadm -u c1t1d0s2 Step 8:Create new metadevice **If you removed metadevice with metaclear command at step 2,you donвЂ™t need this step. #installboot /usr/platform/sun4u/lib/fs/ufs/bootblk /dev/rdsk/c1t1d0s0 Step 7:Update device ID in SVM database #metadb -afc 3 /dev/rdsk/c1t1d0s7 Step 6:Install boot block on вЂќ 0вЂі slice of the new disk. #prtvtoc /dev/rdsk/c1t2d0s2 | fmthard -s вЂ“ /dev/rdsk/c1t1d0s2 Step 5:Add new disk to Metadb

#devfsadm -Cv Step 4:Create Vtoc table for the new diskДЄfter new disk is dedected you need to add vtoc table which is same as mirror disk.If вЂњprtvtocвЂќ command failed use вЂњformatвЂќ to define vtoc table for the new disk. #metaclear d11 Step 3:Add new disk then scan to identify the new disk #metadetach d10 d11 Step 2:Check metadb and remove db replicas which related to failed disk.Д«efore replace broken disk we recommend to remove metadevices.

You вЂll get information about which disk is maintenance state then you need to remove the sub mirror in maintenance state.Use вЂњmetadetachвЂќ command. Not that failed disk is under a mirrored configuration.Д2 s 74GB c1t2d0s0 Step 1:Detach device maintenance state Step 7:Update device ID in SVM databaseДЄt this post, you will find some of the examples of replacing failed disk under SolarisВ volume manager.Step 6:Install boot block on вЂќ 0вЂі slice of the new disk.Step 4:Create Vtoc table for the new disk.Step 3:Add new disk then scan to identify the new disk.Step 2:Check metadb and remove db replicas which related to failed disk.

Root mirroring in veritas volume manager