The goal of high availability (HA) systems is to make system failures tolerable. To implement high availability, the solidDB® HotStandby component enables a secondary server (a standby server) to run in parallel with the primary server (an active server) and keep an up-to-date copy of the data in the primary server.
The HotStandby component implements an internal state machine that makes the server aware of the HA state. The HA state machine makes it possible to preserve database consistency. For example, when the server is in the Secondary state (that is, receiving a transaction stream from the Primary), updates to the Secondary database are disabled.
The availability or states of the servers can be controlled with a watchdog program. The solidDB® server provides a watchdog implementation called the solidDB® High Availability Controller (HAC).
Internally, HAC uses a set of HotStandby commands (HotStandby API) to control the server states. A solution such as this allows implementation of systems that have increased reliability. A failed database server no longer brings your site to a complete halt. In as little as a few hundred milliseconds, in any engine configuration supported by solidDB® (such as solidDB® advance replication with master and replica servers), HotStandby allows the secondary database to replace the failed one.
HotStandby architecture
HotStandby includes the following components:
▪Watchdog application (for example, HAC)
▪HotStandby API (HSB admin commands)
▪Primary and Secondary solidDB® servers
Principles of operation
HotStandby (HSB) performs synchronous transaction replication between two nodes: a primary server and a secondary server.
The primary server node (Primary) contains the active database. The secondary server node (Secondary) contains an exact, up-to-date copy of the active database, and it can replace the Primary if the Primary fails.
The Secondary receives updates from the primary server, and is ready to take over as the Primary if the original Primary fails. An additional benefit of having the Secondary is that the Secondary can also respond to read-only requests (for example, SELECT statements) from clients. This allows you to spread some of your workload over two servers rather than one.
Note The term “hot standby” (two words, all lower case) refers to the general technique of having a second server ready to take over if the first server fails. “HotStandby” (one word, capitalized as shown) refers to solidDB®’s specific implementation of this general technique. The abbreviation for HotStandby is HSB.
Similarly, a watchdog refers to a technology that supervises the state of two databases and can switch the states, if necessary. HAC is the solidDB® watchdog implementation. You can also use the HSB API to program your own watchdog application. The solidDB® package also includes a watchdog sample that you can use as a starting point.
