Recovering from Application Problems

I l @ ve RuBoard

Some application problems can be caused by accidental or inappropriate changes to application configuration files. EMS provides the capability to monitor these configuration files, and can report events whenever a change occurs. The text of the file can be included along with the event, which then can be used to provide a history of configuration changes so that an incorrect change can quickly be undone.

When a server has failed and can't be immediately repaired, high availability cluster software, such as MC/ServiceGuard, can be used to reduce the downtime associated with the failure and keep services available. MC/ServiceGuard can detect the failure of an application and automatically restart the application on another system in under one minute. Some preconfigured solutions already exist for some important applications, such as BEA Tuxedo, CICS, and NFS.

BMC's Application Service Assurance product can be used to provide application recovery. All the components of an application can be recovered to a specific point in time. This product supports many applications, including SAP R/3, PeopleSoft, Baan, and Oracle.

However, even with the kernel's ability to mask certain failure and the high availability software's ability to move applications to redundant servers, ultimately, you still need to repair failed components. For hardware problems, Support Tool Manager (STM) can provide fast diagnosis on HP-UX systems. SyMON can be used for Solaris environments. To fix OS problems, system vendors typically provide support information on their respective Web sites.

I l @ ve RuBoard