Data staging is a new feature in JasperReports Server 6.0 that can optimize the performance of certain Ad Hoc views and reports. With data staging, the entire dataset for a Domain Topic is indefinitely cached in the server's Ad Hoc cache. When the entire dataset is pre-loaded in the server, Ad Hoc views and reports from that Domain Topic run faster because they fetch data from the local cache, not by querying the production database. To keep data fresh, you can specify a reloading interval so that the server periodically accesses the database in the background to reload the entire dataset.
By default, data staging is not enabled on the server. You must enable data staging first, then each creator of a Domain Topic can choose to configure it for their Domain Topic.
When a Domain Topic uses data staging, it no longer uses any Domain-based optimizations (data policies) configured in the server because those apply only to database queries. Datasets that are staged are also exempt from the Ad Hoc cache time limits such as idle time and time-to-live. Data policies for Domain-based optimizations and time limits configured for the Ad Hoc cache still apply to Domain Topics that do not use data staging.
When deciding to enable data staging, you should consider the nature of your data, its size, your database performance, and your access patterns to determine whether staging will help your views and reports run faster. You should also perform realistic usage tests to determine the amount of memory and size of data that give you the best performance. You can see staged datasets on the Ad Hoc cache page (Manage > Server Settings > Ad Hoc Cache), which displays their size and fetch time.
To get the full benefit of data staging, you must ensure your server has enough physical memory allocated to the Java virtual machine so that the cached dataset stays in memory. The cache for data staging will use disk storage when memory is full, but response times will suffer. Therefore, you must apply data staging to carefully selected Domain Topics so that the resulting sum of all staged datasets does not overwhelm your cache or your Java virtual machine (JVM) memory limits.
The cache for data staging also uses disk storage so that staged datasets are persistent during server restarts. Another advantage of staged datasets is that they allow fast reporting even when your original database has latency or downtime issues.
Global Data Staging Configuration
The following setting determines whether the data staging feature is available on the server:
Topic-level Data Staging Configuration
To use data staging, you must first create a Domain Topic by creating an Ad Hoc view and selecting a Domain as its source. After selecting a Domain, use the Data Chooser dialog to specify the sets and fields of the Domain in your Topic, along with any pre-filters and display names. The items chosen from the Domain, the filters in the Domain, along with any pre-filters in the Domain Topic all determine the dataset that data staging will store in the cache.
When data staging is enabled on the server as shown above, Domain Topics offer two new settings in the Save as Topic tab of the Data Chooser dialog.
Data Staging Options in Domain Topics |
The two settings control data staging behavior:
• | Enable staging - Turns on data staging for this Domain Topic. All Ad Hoc editor actions, Ad Hoc views, and Ad Hoc reports based on this Domain Topic will use the staged data in the cache. |
• | Time to keep query result in cache (in minutes) - This value determines the refresh interval in minutes of the staged data. At the end of every interval, the server refreshes the staged data by querying the data source for the entire dataset again. The new staged data replaces the old staged data in the cache so that the staging lasts indefinitely, but the dataset is renewed. Staged data is refreshed asynchronously every interval by the server, independently of any access to the staged data by an Ad Hoc view or report. |
Data Staging Dependencies
Data staging stores a dataset based on the items and filters selected in the Domain and Domain Topic. Views and reports based on the Domain Topic rely on this data being staged, that is, available in the cache.
When data staging has been turned on and views or reports have been created based on this Domain Topic, you can edit the Domain Topic as follows:
• | You can add items from the Domain to the Domain Topic. When you save the Domain Topic, the server immediately retrieves the new dataset with the new items and stores it in the cache. |
• | You cannot remove items from the Domain Topic. If you try to remove items, the Data Chooser dialog will give an error. Look in the log file to find details about these dependencies. |
• | You cannot turn off data staging for this Domain Topic unless all Ad Hoc views and reports based on it have been deleted. |
• | You can change the refresh interval of the staged data. When you save the Domain Topic, the server immediately retrieves the dataset again and applies the new refresh interval in the future. |
Recommended Comments
There are no comments to display.