Archive:The DataStore 4.0: Difference between revisions

From SEPsesam
No edit summary
mNo edit summary
Line 1: Line 1:
{{Template:Navigator_4.0|Topology|Loader}}
{{Template:Navigator_4.0|Topology|Loader}}


===DataStores===
===DataStores===

Revision as of 10:30, 7 June 2011

<<<Back
Topology
User Manual
Next>>>
Loader


DataStores

Until now media types for disk backups (DiskHard and DiskChange) were based on the historic media concept of tape backups. This means a media pool based on these media types has to be separated into logical media.

With the version 4.0.3.x, DataStore has become another milestone in SEP sesam's data management on disk. With it the defined disks are managed as a whole.

Advantages of DataStore:

  • partitions are treated as a unit
  • locking times are defined on a saveset level as opposed to media level
  • dynamic storage sharing

Function

In a classic media pool based on DiskHard the directory of the storage space is defined in the media pool.


DataStore Principle 403a.jpg


With DataStore this definition happens on the drive. As shown in the illustration, a media pool still points to a drive group. Only now one or more DataStores are set in between as an additional level. The connection of a DataStore to the drive is static.

Several DataStores are only needed in a media pool if a media pool is supposed to use the data of several disk partitions. In this case all drives of the DataStores of a mediapool must be part of a drive group. This ensures that the SEP sesam queue manager can distribute the backups on this media pool on all DataStores (Balancing).

The disk space usage management of the DataStore only uses operating system functions of partitions. This means that several DataStores can be set up on a partition, however every DataStore sees the values of the other partitions when checking the partition allocation. This means that the DataStores obstruct each other. It is therefore recommended that you only use one DataStore for each hard disk partition.

Watermarks and Purge

So-called watermarks are available for managing the disk space usage. The High Watermark (HWM) defines how full a DataStore can get until a Purge deletes obsolete files (EOL free). The Low Watermark (LWM) defines how much of the EOL-free files can be deleted from the DataStore. If a LWM=0 is set, all EOL-free savesets are physically removed from the DataStore regardless the filler loading. This value can be used to control how much storage space is made available for files with expired EOLs in the DataStore. This way a data history that exceeds a necessary minimum can be provided on disk. SEP sesam always deletes the oldest savesets first.


DataStore Watermarks-Purge 403a.jpg


Events that lead to a Purge in a DataStore:

  • New day
  • After sharing the drive of the DataStore after a backup
  • Manual Purge in the GUI

Instructions

I. How to set up a DataStore

II. How to use a DataStore

III. How to administer a DataStore

IV. How to maintain a DataStore


DataStore GUI-View 403en.jpg