This topic describes the Delphix Engine performance reservoir and capacity threshold warnings, and various ways to obtain information about capacity and resource usage for the Delphix Sever.
The Performance Reservoir and Capacity Threshold Warnings
- When 78% of the total storage quota is reached, a Warning fault is triggered. You can resolve this fault by deleting objects in the Delphix Engine, adding storage, or changing policies, as described in the topics Adding and Expanding Storage Devices, Deleting Objects to Increase Capacity, and Changing Snapshot Retention to Increase Capacity. Additionally, refreshing target databases will clear space the engine uses to track changes over time for each DB.
- When 80% of the total storage quota is reached, performance may decrease as the ZFS file system begins to leverage an algorithm which prioritizes space-savings over performance for storing data.
- When 85% of the total storage quota is reached, a Critical fault is triggered, and the Delphix Engine will enter Maintenance Mode. When this occurs:
- All pending link, sync, refresh, and provisioning processes will be cancelled, and no new operations can be initiated
- Scheduled replication processes will first check for capacity on the target engine and hold data (replication currently in progress to the engine will not be halted)
- Policy operations such as SnapSync, snapshot, and refresh are suspended for all platforms
- dSources stop pulling in new changes. LogSync is suspended for all Oracle and PostgreSQL dSources. Validated sync is disabled for SQL Server dSources.
- No virtual database (VDB) snapshots can be taken
- When 95% of the total storage quota is reached, a second Critical fault is triggered for SQL Server. All SQL Server VDBs stop in order to maintain data integrity.
To take the system out of Maintenance Mode, increase the storage capacity. You can do this by adding storage devices, deleting objects, or changing policy settings. When you have increased the storage capacity, the system will automatically exit Maintenance Mode.
- When the system falls below 95% of the total storage quota, you can manually start SQL Server VDBs that had stopped
- When the system falls below 90% of the total storage quota, SQL Server VDBs that had stopped will automatically start
- When the system falls below 82% of the total storage quota:
- New link, sync, refresh and provisioning operations are allowed
- Policy operations such as SnapSync, Snapshot, and Refresh resume for all platforms
- dSources start pulling in new changes from their corresponding data sources. LogSync is resumed for Oracle and PostgreSQL dSources. Validated sync is enabled for SQL Server dSources.
For more information, see Setting Quotas.
Ways to View Capacity Usage
You can access capacity and performance information for the Delphix Engine through several different means, including the TimeFlow view, the Dashboard view, and the Capacity screen.
The TimeFlow View
The TimeFlow view provides three summary performance metrics:
- VDBS - all VDBs on the Delphix Engine. Inactive VDBs are greyed out.
- TimeFlow Ratio - the total of Delphix Engine storage against the projected usage in a traditional physical database
- Consolidation Ratio - the amount of space that dSources and VDBs occupy compared to the amount that would be occupied by a traditional physical database
The Dashboard View
The Dashboard view provides more detailed information about the overall performance of the Delphix Engine and its network in five panels:
- Job History - all jobs that have been initiated in the system and their outcome
- Capacity Management - the amount of physical storage available and what is currently used
- TimeFlow Ratio - see above
- VDB Ratio - a measure of the amount of physical space that would be occupied by the database content against the amount of storage occupied by that same database content as VDBs.
- Performance Management - the amount of network bandwidth available and the amount that VDBs are currently utilizing, as well as information about specific VDB network usage
The Capacity Screen
The Capacity screen provides a view of storage allocation for dSources, VDBs, and Snapshots by group and objects within the group, as well as a summary metric. You can also access a graph view of the Capacity screen by clicking Graph View in the upper right corner of the screen. This view shows Available Space as green, Used Capacity as blue, and Reserved Space as yellow.
|Name||Name of the group or database object. Click the expand icon next to a group name to see the objects in that group.|
|Quota||The maximum amount of storage space allocated to the group or object, also known as the ceiling. See Setting Quotas for more information. You can see quota allocations for groups and objects in the Graph view of the Capacity screen.|
|Used||Amount of storage space used by the group or object.|
|Unvirtualized||Estimated amount of space that the group or object would occupy in an unvirtualized state.|
|Ratio||The amount of storage space occupied by the group or object in the unvirtualized state as opposed to the amount of space it occupies as a virtual object. This can also be thought of as the "de-duplication ratio."|
|Keep Until||For Snapshots, the number of days it is retained as set by the Snapshot Retention Policy. See the topics under Managing Policies for more information.|
Viewing Snapshot Capacity Usage
On the Capacity Management screen, select the drop-down icon next to a dSource.
- Select the drop-down icon next to Snapshots. As shown in the screenshot below users can see the space used by snapshots as a whole and by each snapshot individually.
Note: The 'used' value represents the disk space that would be freed if the snapshot was deleted.