	US 7,464,303 B2
	Autonomically adjusting configuration parameters for a server when a different server fails
Eric Lawrence Barsness, Pine Island, Minn. (US); Mahdad Majd, Rochester, Minn. (US); and John Matthew Santosuosso, Rochester, Minn. (US)
Assigned to International Business Machines Corporation, New York, N.Y. (US)
Filed on Jun. 09, 2005, as Appl. No. 11/149,490.
Prior Publication US 2006/0294207 A1, Dec. 28, 2006
Int. Cl. G06F 11/00 (2006.01)

U.S. Cl. 714—48

1 Claim

1. A computer-implemented method for improving performance when one of a plurality of server computer systems fails in a networked computer system that includes a load balancer that routes requests to the plurality of server computer systems, the method comprising the steps of:

detecting failure of one of the plurality of server computer systems;

sending a message indicating the detected failure and how much additional work may be expected due to the failure of the one server computer system, to at least one of the plurality of server computer systems;

measuring a first performance of the at least one server computer system;

the at least one of the plurality of server computer systems autonomically adjusting at least one configuration parameter when the message is received indicating failure of the one server computer system, wherein the at least one configuration parameter includes:

a maximum number of database connections;

a maximum number of prepared statements in a cache; and

a maximum number of threads;

measuring a second performance of the at least one server computer system;

determining whether the adjustment of the at least one configuration parameter improved performance of the at least one server computer system by comparing the first performance with the second performance;

detecting a failure recovery;

sending a message to at least one of the plurality of server computer systems when the failure recovery is detected; and

at least one of the plurality of server computer systems, in response to the message indicating a failure recovery, autonomically adjusting the at least one configuration parameter to normal levels.