Comparison of Database Monitoring Products

I l @ ve RuBoard

Your choice of a database monitoring product is likely to depend on your database responsibilities and the details of your computing environment. For example, EMS is available only for HP-UX environments and has special capabilities if used on servers in an MC/ServiceGuard cluster. The HA cluster products, such as MC/ServiceGuard and Solstice HA, are only available on a single OS platform. Other database monitoring tools, such as MIB Browsers, SPIs, and BMC PATROL, are available for a variety of UNIX platforms.

For configuration management, Network General's Database Module product can help detect when database connectivity problems are caused by misconfigurations. EMS provides a file status change monitor that can be used to ensure that nobody is altering a database configuration file unexpectedly.

For fault management, you can protect your database with either Sun's Solstice HA or Hewlett-Packard's MC/ServiceGuard product, depending on the platform you are using. If MC/ServiceGuard is being used, you can use ClusterView to see graphically any database failures that occur. EMS HA Monitors can provide additional database and server status information. In addition to supporting a variety of notification methods , EMS is the only tool that can send failure information to MC/ServiceGuard. EMS can also send alarms to HP Predictive Support for proactive care of the system, but currently, Predictive monitors only hardware events.

The PATROL KMs from BMC and SPIs from HP provide the most complete solutions for database resource and performance management. Both provide several performance metrics and have solutions for multiple databases. SPIs can be used with MeasureWare and PerfView and are also tightly integrated with IT/O. Because of their tight integration with the IT/O agent, SPIs can provide more sophisticated event correlation and filtering than BMC PATROL. SPIs also have more resource granularity, enabling you to choose the database objects to monitor. For example, you don't need to include read-only tablespaces when monitoring tablespace free space.

For the budget-conscious, EMS HA Monitors can provide some resource monitoring of tablespaces. You can get a lot of useful performance data by browsing the database MIBs, if you have the time. The MIB browsing capability may be useful in small database environments. Operators must bring up the MIB Browser and poll for individual database values, instead of responding to problems on an exception basis. This quickly becomes labor- intensive for even medium- sized installations. However, it can be a useful mechanism for getting additional information after a database problem is reported . MIB data can help diagnose the causes of performance problems. For example, you can go to the Informix Private MIB to see the last SQL query executed by a session.

Thus, a variety of ways to obtain database performance information are available. GlancePlus and MeasureWare add the ability to look at transaction-level performance data. The HA cluster products provide the ability to detect and recover from database problems without user intervention. For example, MC/ServiceGuard can restart a database server on another system in the event of a system failure. BMC PATROL and PLATINUM ProVision provide database administration tools in addition to detection capabilities. SMART Plug-Ins provide textual guid ance on how to recover from problems, but rely on their integration with database vendor tools, such as OEM, for administrative tasks .

For database security and backup management, the tools are pretty weak. Network General's Database Module can detect attempts to breach security because it monitors all database network traffic. IT/O can detect failed superuser logins at the system level. Informix stores some information about past backups in its Private MIB.

Combinations of these tools may be used. For example, BMC PATROL, SPIs, and EMS can each send events to IT/O. Thus, a BMC PATROL database KM can detect a problem and send a message to IT/O's Message Browser, on which an IT/O administrator can already have enabled an automatic action to be taken.

EMS, BMC PATROL, and SPIs each provide additional solutions for system, disk, and network monitoring. Your choice for database monitoring may depend on the products that are already being used in your environment.

I l @ ve RuBoard


UNIX Fault Management. A Guide for System Administrators
UNIX Fault Management: A Guide for System Administrators
ISBN: 013026525X
EAN: 2147483647
Year: 1999
Pages: 90

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net