Slow Historic Display Performance due to Cache Size or Volume of Historic Data
Historic displays such as the Historic List and Historic Trends can suffer from slow performance when:
- The amount of historic data required for the display is too large for the Historic cache
- The volume of historic data is too great as the points’ historic data is being recorded too frequently
- The server hardware is insufficient for the size of database and the amount of points with historic data.
If you are experiencing slow performance with historic displays, you should:
- Determine the number of historic records required for your search.
- If you are displaying a Historic Trend, configure the Key to display the Count, then examine the Count entries. These are the total number of historic records for each point on the Trend. Add up the Count entries for each point shown on the Trend to determine the total number of historic records required for the Trend.
- If you are displaying a Historic List, examine the Rows entry shown in the status bar at the bottom of the ViewX display. You will see that it states the number of rows currently on display and the total number of rows. The total number of rows is the total number of historic records required for the Historic List.
- Repeat step 1 on each unique Historic Search that is active on your system. You need to do this so that you can calculate the total number of historic data records required for all active historic searches on your database (for example, you also need to determine how many historic records are needed by Historic Lists and Trends being used by other users). You only need to include those Historic Lists and Trends that are performing a unique Historic Search—if users are searching the same Historic Records, they will share the same set of records in the cache and so these records have to be counted only once.
- Calculate the total amount of historic records that are required for all active searches.
- Multiply the total amount of historic records by 32. 32 is the number of bytes per historic record. The result of this calculation is the amount of bytes required for the active Historic Searches.
- Compare the total amount of bytes required for the active Historic Searches to the Cache Size for Historic data.
If the Cache Size is sufficient, you may need to reduce the frequency with which historic values are recorded. Proceed to step 7.
If the Cache Size is insufficient, you may need to increase it.
- On 32 bit systems, the total size of the Event Journal and Historic caches combined cannot exceed 400MB and so there is a limit to the amount by which you can increase the cache size (see Define the Historic Data Cache Size).
- On 64 bit systems, you can increase the Historic cache to any size as long as the server has the appropriate amount of RAM available. If there is insufficient RAM, you can upgrade the server to have increased RAM, lower the amount of historic data being recorded or reduce the size of the search being performed on your system.
If increasing the Historic Cache Size does not improve performance, or you are unable to increase the Cache Size to the required amount, proceed to step 6.
- Check that your server hardware meets the system requirements (see ClearSCADA Server Requirements in the ClearSCADA Installation Guide). You may need to upgrade your server or look at ways of reducing the size of your database.
If your server hardware meets the system requirements, proceed to step 7.
- You need to reduce the number of historic values being recorded or reduce the size of the searches being performed (see Excessive Number of Historic Records).
Further Information
Trends: see Using Trends in the ClearSCADA Guide to Trends.