Backup and Recovery - Oracle Exadata Recipes: A Problem-Solution Approach - page 218

How It Works
When patching Exadata storage cells, the patch process uses the active and inactive partitions to provide rollback
and recovery in the event of a patch failure. To see your active and inactive partitions, run the following
imageinfo command:

[root@cm01cel02 ~]# imageinfo | grep device
Active system partition on device: /dev/md6
Active software partition on device: /dev/md8
Inactive system partition on device: /dev/md5
Inactive software partition on device: /dev/md7
[root@cm01cel02 ~]#

In this listing, we see that for this storage cell, our active partitions reside on /dev/md6 and /dev/md8 and the
inactive partitions on /dev/md5 and / dev/md7 .
Exadata storage cell patches can be either Out-of-Partition or In-Partition patches. With Out-of-Partition patches,
the inactive partitions are activated and patched first. When complete, Oracle's validation framework is executed at
reboot and the previously inactive partitions are switched to the active partitions. The previously active partitions are
marked inactive and synchronized with the newly activated partitions.
If during system startup or cell service startup Oracle's validation framework detects an error or failure condition,
the patched partitions are rolled back and the system boots from the un-patched active partitions.
We will cover more details on the Exadata storage cell patching process in Recipes 11-2 and 11-3.

8-11. Recovering Compute Server Using LVM Snapshots
Problem
You wish to recover files or software on your compute nodes using backups taken with LVM snapshots.

Solution
In this recipe, we will assume you have backed up your compute nodes using LVM snapshots as presented in
Recipe 8-5 and walk you through the steps to restore components of your volumes using these LVM snapshot backups.
The steps provided in this recipe will include the following:

• Performing a few pre-recovery checks on your compute node

• Staging your backup image to an NFS mount directory accessible from the failed compute

node on your network

• Attaching the diagnostics ISO file to your CD-ROM drive via the ILOM

• Booting your failed compute node from the diagnostics ISO image

• Performing your restore from the backup image stored on the NFS mount

• Detaching the diagnostics ISO and rebooting the system

• Validating your restore and starting your databases