Bare Metal Recovery for Linux

Overview

Bare Metal Recovery is a disk-based disaster recovery feature that enables administrators to restore an entire Linux system environment for an individual computer. A full recovery using Bare Metal Recovery includes the Linux operating system, its point-in-time backed-up data, and its Oracle applications.

The specific process for using Bare Metal Recovery to restore a machine with a Linux operating system is discussed in this topic and related topics. This involves booting a bare machine, accessing the storage system that contains Bare Metal Recovery backups, and selecting a specific backup instance for restoration. The procedure uses a Linux CentOS-based CD or ISO image.

See also. For the latest system requirements and compatibility details regarding supported hardware, file systems, applications, operating systems, and service packs, go to the DPX 4.11 Compatibility Matrix.

Tip. It is always a good practice to collect logs after a Bare Metal Recovery or to do so after a failure, error, or problem in the Bare Metal Recovery process.

Requirements for Recovering a Linux Machine

Bare Metal Recovery for Linux requires:

The Linux Bare Metal Recovery ISO image file.
A bare machine.
For Linux, the source node must have at least 2 GB of memory.
Information about the storage system that stores the Bare Metal Recovery backups. The administrator recovering with Bare Metal Recovery must have root access privilege on the storage system.
Information about the backup job and the node to recover.

Detailed Considerations

The following are considerations for recovery of a Linux machine:

Bare Metal Recovery can recover data to a dissimilar machine. Because of the large variety and combinations of hardware on servers, that is, vendors, controllers, and processors, it is important to perform careful advanced testing to assess Bare Metal Recovery to dissimilar machines in a given environment.
Most required network drivers, SCSI drivers, or hardware drivers are on the Bare Metal Recovery ISO image file.
When restoring a RHEL 5.x server to a VMware virtual machine using Bare Metal Recovery, the virtual machine must be configured with a supported SCSI adapter, read the Knowledge Base article 46430.
Bare Metal Recovery for Linux supports recovery to physical or VMware virtual machines. Perform careful advanced testing to ensure required recovery in a given VM environment. There are some combinations of dissimilar hardware restore where the Linux vendor’s distribution will not automatically enable the correct drivers for storage, video, and networking. In these cases, it is strongly advised that you have a copy of your vendor’s install/rescue media available to correct your booting system after the Bare Metal Recovery recovery is completed.
Bare Metal Recovery restores from vStor are supported for Linux Kernel 3.10 and later.
The Bare Metal Recovery process is not supported for use with software RAID or fake RAID.

The following are considerations for Bare Metal Recovery for Multipath devices on Linux:

Consult the DPX Compatibility Guide to determine which versions of Linux support Multipath devices withDPX Bare Metal Recovery.
To perform a Bare Metal Recovery where the whole original source machine was on a Multipath device,ensure iSCSI SAN BOOT is configured properly.
To perform a Bare Metal Recovery to the original source machine, where the original source machine wason a Multipath device, you must power off the original source machine before starting the recovery.

Recovery Procedure on Linux

Preliminary Steps

Before running a recovery procedure, make sure the following steps are completed.

Schedule and run Bare Metal Recovery backup. See Running Backups.
Track information about the run Bare Metal Recovery backups. See Tracking Important Information.
Set up the bare machine. See Bare Machine Considerations.
Ensure all requirements for Bare Metal Recovery for Windows are met. See Requirements for Recovering a Windows Machine.

Booting on Linux

Boot from the Bare Metal Recovery disc. While the machine is booting, diagnostic messages may appear.

To boot from the disc:

The procedure for booting from a disc varies, but for most scenarios, adjust the BIOS configuration for the correct sequence of bootable devices. In many cases, when the computer is turned on, a message describes how to change the BIOS configuration.

After the bare machine boots up the Bare Metal Recovery CD successfully, all the modules are loaded into the machine’s RAM.

To relaunch the application without rebooting:

Launch Bare Metal Recovery with the following script based on the access to distribution version:

/root/NSB/scripts/bmr.sh (for BMR-<xxx>-Linux-2.6.iso, where <xxx> is the version)

/root/DPX/scripts/bmr.sh (for BMR-<xxx>-Linux-3.10.iso, where <xxx> is the version)

Running the Recovery on Linux

Note. If there is a failure, error, or problem at any step in the Bare Metal Recovery process, click the Logs button to collect the diagnostic logs and attach them to a customizable email message.

To recover on Linux:

Boot up the bare machine using the Bare Metal Recovery disc. Because the bare machine serves as the restore destination, the bare machine must have network access to the storage system where the backup images reside. Note that access control on the storage system may need to be modified to allow restoration to the bare machine.
Configure the appropriate network settings for the new bare metal machine as needed. The Storage System dialog appears.
In the storage system group, enter the requested information as follows:
IP Address
The IP address or fully qualified hostname of the storage system that contains the required backup instances.
Username
The User ID for logging into the storage system. Use the root account for full access to all data resources.
Password
The password for logging into the storage system with the User ID. Example: syspass.
Type
The type of storage from which you are recovering. Choices include NetApp, Catalogic vStor, or Catalogic OSS (DPX Open Storage Server).

Tip. Data input is case-sensitive for these fields.

Click Connect once all the information is entered accurately.
Enter the requested information as follows:
Volume Name
Lists the name of each volume on the storage system that contains backup snapshots. Select a volume name.
Node Name
Lists the name of each node that has one or more volume backups stored on the storage system. Select a node to restore.
Job Name
Lists the name of each backup job stored on the storage system for this node. A change of the Node Name selection automatically repopulates the selections offered in the Job Name list. Select a backup job name.
Select a snapshot to restore from the Select a Snapshot to Restore group, which lists all of the snapshots in the storage system that are available for this node and job selection. Each job in the list has a unique date and time stamp.
Click Next.
Select the disks to partition, then map restore source disks to bare machine target disks:
1. Use checkboxes to select source disks for partitioning. Note that the check box for the root boot directory cannot be cleared.
2. Reconfigure the disk mappings if necessary. To change a target disk selection, click the name of a Target Disk listing to reveal a list of alternate target disk options.
3. Click Next. The Select Source Drives to Restore dialog appears.
On the Select Source Drives to Restore dialog, note the following:
Only the selected drives, as indicated with a check mark, are recovered as part of this Bare Metal Recovery job.
In the Type field, the letter B identifies the boot volume and the letter S identifies the system volume. System and boot volumes are default selections that are mandatory and unchangeable.
When performing a partial-node restore, consider restoring the drive that contains DPX, which eliminates the need to reinstall DPX.
Use an ordinary Block restore to recover a non-system drive not recovered through Bare Metal Recovery. See Block Restore.
Change drive selections if necessary. Click Next to partition the disks on the bare machine.
A warning message appears: “This action will overwrite the data on the target drives. This action cannot be reversed. Do you wish to continue?”
Click Yes to continue, click No to return to the previous window and modify your selections.
The summary dialog box appears.
Note that a dash that appears next to a drive mapping indicates that Bare Metal Recovery did not create this particular partition. If a source volume was mapped to this partition in a previous step, Bare Metal Recovery is not able to restore this data. This includes any logical volumes that may exist on this target drive.
By default, this process automatically installs the full set of available drivers for the recovered machine.
Click Restore.
For Red Hat 7.x only, special steps are needed due to a new network device naming convention for RedHat 7.x. Catalogic provides a script that helps you manage the NIC addressing on the BMR target machine. Before continuing to the next step, use SCP (secure copy) or a similar tool to copy this script from the ISO to different accessible location, such as a USB or network drive making note of the correct path based on distribution version:
/root/DPX/scripts/Post_NIC_Config.sh (for BMR-<xxx>-Linux-3.10.iso, where <xxx> is the version)
/root/DPX/scripts/Post_NIC_Config.sh (for BMR-<xxx>-Linux-4.18.iso, where <xxx> is the version)
Make sure you copy this script from the ISO image to the different location prior to rebooting the BMR target machine, which occurs in the next steps. However, do not run the script until the BMR target machine has been rebooted.
If Bare Metal Recovery is not able to create a particular partition, a warning message appears: “Bare MetalRecovery could not prepare the target for recovery. Manually remove all existing target partitions and reboot this machine before restarting this process.” Click OK to continue.
After recovery completes, a message appears indicating that drivers were installed successfully or summarizing missing drivers. Click OK.
Bare Metal Recovery applies the configuration changes if needed and reboots the machine. After reboot,the machine contains the newly recovered environment and data.
For Red Hat 7.x only, manage the NIC addressing conventions by either of these methods:
- Manually by using Linux network configuration commands (such as ifconfig), or
- By running the Post_NIC_Config.sh script you copied in Step 15. The script enables you to assign the NIC IP address, subnet mask, and gateway IP address.

Logs Collection for Linux

If there is a failure, error, or problem at any step in the Bare Metal Recovery process, click the Logs button to collect the diagnostic logs and attach them to a customizable e-mail message.

To capture and transmit diagnostic logs:

On any dialog box in the Bare Metal Recovery process, click Logs. The ExpressMail dialog appears.
Enter the Sender/Recipient information and Server/Authentication information as requested.

Note. The To: field should contain the email address of the person who handles the diagnostic troubleshooting. To send diagnostic logs to multiple recipients, separate their email addresses with semicolons.

Clear any check boxes that do not apply.
1. If the SMTP server does not support user name and password authentication, clear the Server Requires Authentication check box.
2. If the SMTP server does not support Transport Layer Security (TLS), clear the Use Secure Connection (TLS) check box.

Note. For a successful log file transmission, the bare machine must be able to access the network when connecting to the specified SMTP server.

Click Send. Bare Metal Recovery generates and accumulates the diagnostic logs, then sends them to the e-mail recipient over the network.

Note. At the end of a recovery, Bare Metal Recovery collects all log files and stores the collection under the root directory based on distribution version:

/root/DPX/DPX_BMRLogs.tar.gz (for BMR-<xxx>-Linux-3.10.iso, where <xxx> is the version)
/root/DPX/DPX_BMRLogs.tar.gz (for BMR-<xxx>-Linux-4.18.iso, where <xxx> is the version)

PreviousBare Metal Recovery for Windows NextRecovery from a Replicated or Alternate Data Resource

Last updated 1 month ago