US 7,502,796 B2 · Arrangement and method for optimizing performance and data safety in a highly available database system

Abstract

A method and arrangement for optimizing performance and data safety in a system includes first and second database servers coupled to a communication network. The first database server accepts transactions from a client application, stores them according to a data storage method, and sends them to the second database server via the communication network, whereupon the second database server also store the transactions. The first database server receives information about the server availability state of the second database server and detects changes in the availability state of the second database server in a such manner, that the data storage method of the first database server is changed according to the change in the availability state of the second database server.

Claims39

Dependent claims

A method for optimizing performance and data safety in a system comprising a first and at least one second database server being coupled with a communication network, comprising the steps of:

the first database server accepting at least one transaction from a client application;
the first database server attempting sending the transaction to the second database server via the communication network to replicate the transaction on the second database server before the transaction is stored in the first database server;
the first database server determining one of i) a success and ii) a non-success, of the sending the transaction to the second database server to replicate the transaction on the second database server;
and
based on said determination of sending the transaction to the second database server, the first database server storing the transaction in a persistent storage according to a data storage method in one of a buffered disk write mode and an unbuffered disk write mode.

2 more elements

An arrangement for optimizing performance and data safety, where the arrangement comprises:

a first and at least one second database server being coupled with a communication network,
the first database server being adapted to accept at least one transaction from a client application, store said transaction according to a data storage method,
said first database server being adapted to send the transaction to the second database server via the communication network and the second database server adapted to receive the data of said transaction to replicate the data of the transaction on the second database server before the data of the transaction is stored in the first database server, wherein:
the first database server is further adapted to i) receive information about the server availability state of the second database server, ii) to detect a change in said availability state of the second database server, and iii) to change the data storage method of the first database server, for storing the data of the transaction in a persistent storage, between buffered disk write and unbuffered disk write modes according to said detected change in the availability state of the second database server.

1 more element

A first database server for optimizing performance and data safety in an arrangement, where the arrangement comprises:

said first and at least one second database server being coupled with a communication network,
the first database server being adapted i) to accept at least tranz one transaction from a client application, ii) to send the transaction to the second database server via the communication network so that the second database server receives the data of said transaction to. replicate the data of the transaction on the second database server before the data of the transaction is stored in the first database server, and iii) to store said transaction in a persistent storage according to a data storage method, the storage being subsequent to send the transaction to the second database server, wherein:
the first database server is further adapted to receive information about the server availability state of the second database server, detect a change in said availability state of the second database server, and change the data storage method of the first database server, to the persistent storage, between buffered disk write and unbuffered disk write modes according to said detected change in the availability state of the second database server.

A computer readable storage medium storing program instructions computer executable to cause a first computer database server to function for optimizing performance and data safety in a system comprising the first database server and at least one second database server coupled with a communication network, where the instructions when run in the first database server, cause the first database server to accept at least one transaction from a client application, store said transaction according to a data storage method, and send the transaction to the second database server via the communication network so that the second database server receives the data of the transaction to replicate the data of the transaction on the second database server before the data of the transaction is stored in persistent storage in the first database server, wherein:

the instructions, when run in the first database server, cause the first database server to i) receive information about the server availability state of the second database server, ii) detect a change in said availability state of the second database server, and iii) change the data storage method of the first database server, for storing the data of the transaction in a persistent storage, between buffered disk write and unbuffered disk write modes according to said detected change in the availability state of the second database server.

A computer readable storage medium storing a watchdog application with program instructions computer executable for optimizing performance and data safety in a system comprising a first database server and at least one second database server coupled with a communication network, the first database server being adapted to accept at least one transaction from a client application, store said transaction in persistent storage according to a data storage method, said first database server being adapted to send the transaction to the second database server via the communication network and the second database server adapted to receive the data of said transaction to replicate the data of the transaction on the second database server before the data of the transaction is stored in the first database server, wherein:

the program instructions of the watchdog application cause the first database server to i) monitor the server availability state of the second database server, ii) detect a change in said availability state of the second database server, and iii) change the data storage method of the first database server, for storing the data of the transaction in a persistent storage, between buffered disk write and unbuffered disk write modes according to said detected change in the availability state of the second database server.

Dependencies are inferred from the claim text.

Arrangement and method for optimizing performance and data safety in a highly available database system

Life of this patent

Arrangement and method for optimizing performance and data safety in a highly available database system

Life of this patent

Abstract

Claims39