Determining whether to use container-copy storage pools for disaster protection

Edit online

Determine whether container-copy storage pools meet your requirements for disaster protection.

About this task

Tip: Alternatively, you can use copy storage rules to automatically create a daily, hydrated copy of your directory-container storage pool to tape for disaster recovery. For information, see Comparing options for copying directory-container storage pool data to tape.
You can create an offsite copy of your container-copy storage pool for disaster recovery protection or to satisfy regulatory and business requirements for offsite tape copies. Before you decide to use offsite tape copies for disaster protection, carefully consider whether the solution meets your recovery time objective.
Using container-copy storage pools for disaster recovery is suitable when the amount of data in your environment is equal to or less than the following values:
  • 200 TB of total managed data
  • 50 TB of back-end data
  • 37 TB of front-end data
Total managed data
All data that is stored in the directory-container storage pool on the server. This includes active and inactive versions of the data. The number of versions is determined by retention policies.
Back-end data
All data that is stored in the container-copy storage pool.
Front-end data
The current active data that is stored in the container-copy storage pool. This is the active data that is used to restore data on client nodes. In a disaster, all or part of the front-end data is required to reestablish production. Front-end data is a percentage of total managed data and is less than or equal to the total managed data, depending on the policy settings in use.
To recover from a disaster within 48 hours, the system environment at the recovery site must meet the minimum hardware requirements for the actions in the following table.
Action Time required Minimum requirements
Configure a new IBM Storage Protect server at a disaster recovery site. To configure the new server, you must complete the following steps:
  1. Provision disks for the server.
  2. Restore the server from backup.
  3. Start the server.
  4. Update the storage and device configurations.
 Time to restore the server: 6 hours Use a solid-state drive (SSD) for the server database, with the following requirements:
  • A minimum of 100 MB per second average combined read/write throughput
  • A minimum of 12,862 average input/output operations per second (IOPS)
Audit the directory-container storage pool and repair the data from tape.
Tip: If the system meets the minimum hardware requirements, you can repair up to 50 TB of back-end data within 48 hours.
 Time to audit the storage pool: 2 hours

Time to repair the storage pool by using a tape copy: 28 hours

Note: The time estimate applies if you have a maximum of 200 TB of total managed data in the storage pool.
 Use Nearline SAS (NL-SAS) drives, as in a medium Blueprint server configuration, with a minimum of 700 MB per second write performance to storage pool disk.

Use new generation tape technology such as LTO-7 or better, with a minimum of six drives to allow concurrent read operations from tape volumes.
Restore data on client nodes.
Tip: If the system meets the minimum hardware requirements, you can restore up to 37 TB of front-end data within 48 hours.
 Time for client restore operations: 12 hours Use NL-SAS drives, as in a medium Blueprint server configuration, with a minimum of 10 restore sessions achieving 3102 GB per hour.

Procedure

  1. Estimate the disaster recovery time for your environment by using the following table. Determine whether the recovery time meets your requirements.
    Table 1. Recovery time estimate for differing amounts of total managed data
    Recovery time objective Total managed data (TB) Number of hours to repair a directory-container storage pool (First Byte Restored) Hours until client nodes are restored (Disaster Recovery complete)
    Up to 1 day 25

    50

    75

    		 10

    13

    17

    		 12

    16

    22
    Up to 2 days 100

    200

    		 20

    34

    		 26

    46
    Up to 4 days 300

    400

    		 48

    62

    		 66

    86
    More than 4 days 500 76 106
    Notes:
    • Achievable rates are highly dependent on the workload and the configured environment.
    • The front-end data percentage is relative to the total managed data. Increasing the amount of front-end data increases the total recovery time. Decreasing the amount of front-end data decreases the total recovery time.
  2. Estimate the recovery time for your environment by using the following formulas:
    • Estimate the value Hours until directory-container storage pool is repaired (First Byte Restored): 
      Time to Client First Byte Restore = 
6 hours + 14 hours for every 100 TB of Total Managed Data
    • Estimate the value Hours until client nodes are restored (Disaster Recovery complete): 
      Time to Client Restore Complete = 
Time to Client First Byte Restore + ((Total Managed Data * Front-End Data) / Restore Rate)

    Restore Rate: The rate at which clients can restore data from the server back to their local computer or storage device.

  3. Complete test procedures for disaster recovery to ensure that container-copy storage pools can be used to restore your environment in a timeframe that meets your requirements.