Oracle Autonomous Health Framework Problem and Solution Space

Oracle Autonomous Health Framework (AHF) maximizes availability and performance by enforcing best practices, capturing data at first failure, monitoring the whole system (server, database, I/O, and network) to proactively discover issues and notify the user and provide timely bug resolution by suggesting fixes automatically after failure.

System administrators can use most of the components in Oracle Autonomous Health Framework interactively during installation, patching, and upgrading. Database administrators can use Oracle Autonomous Health Framework to diagnose operational runtime issues and mitigate the impact of these issues.

Availability Issues

Availability issues are runtime conditions that threaten the availability of the software stack.

These issues can originate from software components, such as Oracle Database, Oracle Grid Infrastructure, and the operating system, or from underlying hardware resources, such as CPU, memory, network, and storage.

Oracle Autonomous Health Framework (AHF) components help detect, diagnose, and resolve availability issues.

Server Availability Issues

Server availability issues can cause a server to be evicted from a cluster, resulting in the shutdown of database instances running on that server.

Examples include:

• Issue: Network congestion on the private interconnect can cause excessive latency or dropped packets for time-critical internode communication or storage I/O. This type of problem can develop gradually and might be detected before it causes an outage.

  Solution: If the issue results from a server configuration change and persists for more than an hour, Cluster Verification Utility (CVU) can detect it. Oracle Cluster Health Advisor can detect the issue within minutes and provide corrective actions.

• Issue: A private interconnect failure caused by a disconnected cable or failed network interface card (NIC) can result in immediate node eviction.

  Solution: Although this type of failure cannot generally be detected in advance, Cluster Health Monitor and Oracle Trace File Analyzer can help isolate the cause by identifying the time of the failure and the network interfaces involved.

Database Availability Issues

Database availability issues can cause an Oracle Database or one of its instances to become unresponsive and unavailable to applications and users.

Examples include:

• Issue: Runaway queries or delays can prevent other sessions from accessing critical database resources, such as locks, latches, or CPU. Resource contention can cause a database or database instance to become unresponsive.

  Solution: Blocker Resolver can detect and automatically resolve these types of delays. Oracle Cluster Health Advisor can also detect and identify the delays, notify the database administrator, and recommend corrective actions.

• Issue: Denial-of-service (DoS) attacks, vulnerabilities, or software defects can cause a database or database instance to become unresponsive.

  Solution: Oracle Orachk provides proactive recommendations for known issues and their resolutions, helping prevent some of these conditions. If an issue occurs, Oracle Trace File Analyzer can automatically collect relevant logs, while Cluster Health Monitor provides additional diagnostic data to help accelerate investigation and resolution.

• Issue: Configuration changes can cause database outages that are difficult to diagnose. For example, incorrect permissions on the oracle.bin file can prevent session processes from being created.

  Solution: Use Cluster Verification Utility and Oracle Orachk to identify and correct configuration-related issues. Oracle Orachk can generate a diff report that compares two reports and identifies configuration changes. You can also use Cluster Verification Utility configuration reports to verify that the system satisfies Oracle installation requirements.

Performance Issues

Performance issues are runtime conditions that adversely affect system performance.

These issues can originate from:

• Software-related conditions, such as software defects, configuration problems, resource contention, or unpatched issues.
• Client-related conditions, such as workload demand, query characteristics, connection management, or resource configuration.

Database server and client performance issues are often closely related. For troubleshooting purposes, they can be categorized according to whether they originate primarily from the database server or the database client.

Database Server Performance Issues

Examples include:

• Issue: Deviations from recommended configuration best practices can degrade database server performance.

  Solution: Oracle Orachk periodically evaluates the environment for configuration issues and notifies the database administrator of recommended corrective settings.

• Issue: A session can block other sessions, causing them to wait for the blocking session to release a resource or complete its work.

  Solution: Blocker Resolver detects chains of blocked sessions and can automatically terminate the root blocker session to relieve the bottleneck.

• Issue: Unresolved known issues or unpatched software defects can adversely affect database server performance.

  Solution: Oracle Orachk can identify known issues during scheduled compliance checks and provide information about applicable patches or workarounds. Oracle Orachk is regularly updated to include newly identified critical issues in existing and new product areas.

Performance Issues Caused by Database Clients

Examples include:

• Issue: Incorrectly configured parameters, such as System Global Area (SGA) and Program Global Area (PGA) allocations, session or process limits, and CPU counts, can degrade database performance.

  Solution: Oracle Orachk can identify configuration settings that deviate from recommendations, while Oracle Cluster Health Advisor can detect the resulting performance effects. These components notify you of detected issues and provide recommended corrective actions.