4 4 3 Grolar:Configuring Si3 Deduplication Store

From SEPsesam
Jump to: navigation, search
Other languages:
Deutsch • ‎English

Copyright © SEP AG 1999-2022. All rights reserved.

Any form of reproduction of the contents or parts of this manual is allowed only with the express written permission from SEP AG. When compiling and designing user documentation SEP AG uses great diligence and attempts to deliver accurate and correct information. However, SEP AG cannot issue a guarantee for the contents of this manual.

Icon archived docs.png This is not the latest version of SEP sesam documentation and, as such, does not provide information on features introduced in the latest release. For more information on SEP sesam releases, see SEP sesam Release Versions. For the latest documentation, check Configuring Si3 NG Deduplication Store.

Information about deduplication store types

SEP sesam v. 5.0.0 Jaglion has introduced a new generation Si3 deduplication store: Si3 NG. The Si3-related information differs slightly depending on which datastore is used: Si3 or Si3 NG. The procedures presented in this article apply only to the older type of SEP Si3 deduplication store, which will soon be obsolete. To learn more about the new generation of Si3 NG, see Encrypting Si3 NG Deduplication Store.

It is strongly recommended to use the newer type SEP Si3 NG deduplication store. It offers significantly higher performance for backup, restore and migration, as well as direct backup to S3 and backup to Azure, the new immutable storage feature SiS, resulting in improved performance, scaling, and resource savings. For details, see the Si3 and Si3 NG comparison section below.
To learn how to upgrade from Si3 to Si3 NG by configuring a new Si3 NG on the same host and creating a replication job to replicate from Si3 to Si3 NG, see Configuring Si3 NG Deduplication Store.

Overview

SEP sesam provides a target-based (Si3T) and source-based deduplication (Si3S). For details on deduplication concept and recommendations, see Deduplication.

Both, Si3T and Si3S require a configured Si3 deduplication store. Only one Si3 deduplication store can be configured on a server. A valid licence is required for each Si3 deduplication store. Note that you can also configure an Si3 deduplication store by using a command line. For details, see Configuring and Administering Si3 Deduplication Store by using CLI.

You can download SEP Tachometer to analyse the structure of your data and calculate potential savings with SEP sesam Si3 deduplication. Check SEP Tachometer.

Prerequisites

For the minimum Si3 hardware requirements that apply to SEP sesam Si3 deduplication server, see Hardware requirements. Keep in mind that these requirements represent the demand for deduplication only. In addition, the amount of memory for the operating system and other services should be taken into account.

In addition, the following prerequisites must be met to configure an Si3 deduplication store.

Additional RAM / CPU requirements

  • For details on the required Java version, see Java Compatibility Matrix. Si3/Si3-NG is not mandatory, so there is no dependency rule for it in the RPM/DEB packages.
  • When estimating the maximum size of a deduplication store, you have to ensure that there is enough space available for dedup trash, otherwise the deduplication store will run out of space. You should calculate the required disk space based on a representative sample of your full backup and add the additional storage space equal to approximately 50% of the representative full backup.

Disk attachment and protocols

Si3/Si3-NG supports all types of direct-attached disk storage, such as serial attached SCSI (SAS), Serial ATA (SATA), and Fibre Channel (FC)/LUN.

Performance tip

Applies to Windows only: SEP AG recommends using the High performance power plan to increase the performance of your backup. Note that Windows sets all computers to the Balanced power plan by default and you must manually switch to the High Performance power plan. This way, your Windows computer will use more power, but the systems with Si3 will always operate at the highest performance level.

  • From the Start menu, go to Control Panel -> System and Security -> Power Options and change the setting to High performance.

Restriction

To avoid problems resulting from the combination of excessively large Si3 deduplication stores and inefficient hardware, the maximum initial Si3/Si3-NG deduplication store size is limited to 40 TB. If you would need to increase this limit, contact SEP support.

This limitation applies to the creation of a new Si3/Si3-NG deduplication store in the GUI.

Information sign.png Note
It is recommended to run Si3 deduplication (SEP sesam Server or RDS) on the physical host. It is also possible to run it on a virtual machine. In this case, take into account that deduplication consumes a lot of server resources for reading, processing and writing the deduplicated data, as well as for some other deduplication tasks such as housekeeping and various checks. These tasks require a large amount of IO and a large amount of memory. Si3 performance can be affected by other virtual machines running on the same host. Therefore, if you are running Si3 on a VM, you should be aware of possible bottlenecks and shortcomings.

Required additional amount of RAM and CPU cores

Memory requirements are dependent on the number of concurrent streams and expected workload. The following tables show the recommended minimum additional amount of RAM and CPU cores for a Si3/Si3-NG data store to ensure good performance. The TB value corresponds to the capacity of the Si3/Si3-NG data store.

Information sign.png Note
These requirements relate solely to the need for deduplication. In addition, you should consider the amount of memory for the operating system and other services.
Si3/Si3-NG data store capacity (check initial size limit) RAM
<20 TB at least 16 GiB
20-40 TB at least 32 GiB

The following table shows the number of CPU cores required for a Si3/Si3-NG data store. The TB value is the amount of data backed up (before deduplication)!

Backed up data (before dedup) CPU cores Note
10 TB 4
20 TB 4
40 TB 8
Note

This is the minimum amount to ensure good performance. Depending on the number of concurrent streams, more cores may be needed.


Steps

The SEP sesam data store is a disk based storage that enables savesets (backed-up data) to be backed up directly to the configured storage locations. SEP Si3 target deduplication is easy to configure and ready to use by selecting the Si3 deduplication store type.

  1. In the Main selection -> Components, click Data Stores to display the data store contents frame.
  2. From the Data Stores menu, select New Data Store. A New Data Store dialog appears.
  3. Under the Data store properties in the Name field, enter a meaningful name for the Si3 data store.
  4. From the Store type drop-down list, select SEP Si3 Deduplication Store.
    SEP Tip.png Tip
    As of SEP sesam v. 5.0.0 Jaglion, two Si3 data store types are available. SEP sesam recommends using the new generation Si3 NG when creating a new data store. To learn how to upgrade from Si3 to Si3 NG by configuring a new Si3 NG on the same host and creating a replication job to replicate from Si3 to Si3 NG, see Configuring Si3 NG Deduplication Store.
  5. 850px

  6. Make sure that the option Create drive is checked under the Drive parameter properties. The predefined value for the drive is automatically inserted in the Drive number field.
  7. It is recommended that the option Create second drive is also enabled. Without it, SEP sesam can only allocate a drive for either reading or writing, running one job on the same drive at a time. By using the additional dedicated drive for restore, you can perform a backup on the first drive and restore your data from the second drive simultaneously. You can also add a third drive for migration.

  8. Then select Create new drive group and enter the name for your Si3 deduplication store dedicated group.
  9. The predefined number of channels is already displayed in the Max. channels drop-down list. The number of available channels depends on your SEP sesam Server package. For details on licensing, see Licensing.
  10. From the Device server drop-down list, select the device server for your data store.
  11. In the Path field, enter the location of your data store or use the Browse button to select the relevant folder. Check the relevant folder and click OK.
    When using the Browse button to select the folder, the New Data Store information window appears with predefined recommended values for your Si3 deduplication store size. Click OK to confirm the selected location and recommended size values. You can modify your data store size later under the Size properties (see step 10).
  12. GUI new data store information Beefalo V2.jpg

  13. Under the Size properties, specify or modify the following:
    • Capacity: Specify the size (in GiB) of the partition for backups.
    • High watermark: Specify the value (in GiB) for the high watermark (HWM). The HWM defines the upper value for the used storage space. When this value is reached, the status of a datastore changes from OK to Warning, but backups continue to be performed. Make sure that you provide enough storage space for your backed up data.
    • Si3 repair area: Specify the value (in GiB) for the Si3 repair area. The Si3 repair area (subdirectory trash) defines space for Si3 files (DDLs) that were identified by a garbage collection job and are no longer used. These files are still kept in the repair area to enable possible repair of Si3 in case of structural problems (which may be caused by a file system error or an operating system crash)). The files in the repair area are automatically removed after the specified period of time (SEP sesam default: 4 days) or when or when the disk usage threshold is reached. Note that the Si3 repair function is disabled when the value is set to 0.
    • Information sign.png Note
      The Si3 repair area for managing disk space dedicated to Si3 files (DDLs) is only available in the Expert UI mode. So if you run the GUI in Basic or Advanced UI mode, you first have to change the mode to Expert, as described in Selecting UI mode.

    The Disk space usage properties are used by SEP sesam to report the following:

    • Used: Total used space (in GiB) on the partition.
    • Total: Maximum available space (in GiB) on the partition as reported by the operating system.
    • Free: Available disk space (in GiB) for SEP sesam.

You can view the properties of your Si3 deduplication by double-clicking the corresponding Si3 deduplication store.

  • You can modify existing and set additional drive options by double-clicking the first drive. In the Drive Properties window, you can browse the path for the data store, set the access mode for data store drives, etc. To configure Si3 data encryption, you have to create a deduplication security password file. For details, see Encrypting Si3 Deduplication Store.
    Use the tab OS Access and specify the credentials to access the respective systems. Use DOMAIN\USER format for domain accounts or HOST\USER for local accounts.
  • Si3 drive properties Beefalo V2.jpg


Monitoring deduplication status

You can view the status of your Si3 deduplication in the GUI by clicking the Si3 State tab. You can check the last deduplication message, the status of active tasks, the encryption status, the number of stored objects, the data size before/after deduplication, the dedup ratio, the saved storage space, etc. Si3 state tab Beefalo V2.jpg

SEP Tip.png Tip
You can also check the status in SEP sesam Web UI. You can use the Dependencies tab to see the diagram of all dependencies of the datastore and drives, RDS, media pools, etc., and use other tabs to check the details of the data store operation.
Information sign.png Note
If fsck (file system consistency check) detects an irregularity in the Si3 file system, the affected pages and chunks are recorded in recovery.log. The Si3 deduplication store in the GUI and Web UIis highlighted in red and Si3 purge is no longer executed. The purge is stopped to prevent the files in the Si3 repair area from being deleted, as they may be needed to repair Si3 in case of problems. Once the errors are fixed and the recovery.log is empty, the Si3 data store is no longer marked in red and the Si3 purge works again.

Si3 deduplication store red Beefalo.jpg

What is next?

After configuring the Si3 deduplication store, configure the media pools first then set up your backup strategy.

Comparison of Si3 and Si3 NG

SEP sesam v. 5.0.0 Jaglion has introduced a new generation Si3 datastore: Si3 NG. It offers significantly higher performance for backup, restore and migration, as well as direct backup to S3, resulting in improved performance, scaling and resource savings. The following table compares the main features of the two deduplication data stores.

Function Si3 Si3 NG
Si3 backup YesY YesY
Si3 deduplication (source-side and target-side) YesY YesY
Si3 replication: local to remote store Notea YesY Si3 to Si3 YesY Si3 NG to Si3 NG
Si3 replication: to S3 cloud YesY YesY (see below Direct to S3)
Si3 NG direct to S3 YesY YesY
SiS (SEP Immutable Storage) YesY (as of Jaglion V2) YesY
Si3 restore YesY YesY
Si3 encryption YesY (as of Jaglion V2) YesY
Seeding Si3 deduplication store Noteb YesY YesY
Usage of tachometer YesY YesY
Notea

*SEP sesam does not support a direct upgrade from the old Si3 to Si3 NG. However, to use the new Si3 NG you can:

  • Back up all data again to the newly configured Si3 NG deduplication store.
  • After configuring a new Si3 NG, you can also create a replication job to replicate from the Si3 to the Si3 NG store. Replication reads all the data from the source-side store on the source-side RDS and sends it to the target store using the source-side deduplication function. For details, see Replicating from Si3 to Si3 NG.
  • You can also configure a new Si3 NG and an old Si3 in parallel on the same host by enabling the key enable_gui_allow_multi_dedup.
Noteb

The Initial Seed feature does not work in v. 5.0.0 Jaglion, but you can use it in earlier SEP sesam versions.