Data Buffering

Data buffering acts as a protection against data loss in case of:

  • a disconnect (overflowing the RAM capacity -> new data overwrites older data in the RAM -> data get lost before reconnecting and fast enough writing data into the DB)

  • a too big data input by one-time, which cannot be buffered in the RAM completely

  • a crash of the RDB manager (data in the RAM get lost)

In order to reduce the risk of a data loss to a minimum, data buffering offers two options, how data blocks (events and alerts) can be saved in your system. In addition to the typical buffer functionality, which saves data blocks in the main memory (RAM), it's also possible to save data on the local hard drive disc (HDD). In case of a database disconnect, the RDB manager is continuously trying to reconnect and takes care that no data get lost. After re-establishing the connection to the database, the buffered data are written into the database (1. RAM, 2. local hard disc).

Individually you can decide from which moment the data should be saved on the HDD:

  • BufferToDiskMin – data blocks are buffered in main memory. As soon as the capacity in the RAM is not enough, data blocks are buffered on the local HDD.

  • BufferToDiskMax – data blocks are buffered from the beginning in the RAM and on the local HDD. As soon as the capacity in the RAM is not enough, data blocks will be saved only on the HDD.

Note:

Note that data buffering had been developed for read access only. To read data from the database without a DB connection is hence impossible!

For more detailed information about the BufferToDisk functionality, please read the following chapters.

Chapter Description
BufferToDisk, basics Introduction and links to the following chapters.
BufferToDiskMin Definition and behavior of the BufferToDiskMin mode.
BufferToDiskMax Definition and behavior of the BufferToDiskMax mode.
Details on data buffering Detailed information on data buffering - write access, limit for data buffering, simulation of an interruption of the database connection, start of an RDB Manager with buffering and without connection to the database.