Search Results ad_task_timing

Overview

AD_TASK_TIMING is a table in the APPLSYS schema owned by the Applications DBA (AD) product. It records execution timing and status information for individual tasks performed by AutoPatch and related AD utilities (adpatch, adadmin, adcfgclone) during Oracle E-Business Suite maintenance operations. Each row represents a single task invocation — such as applying a patch driver, running a SQL script, generating a form, or compiling a Java class — along with the session, worker, timing, and status context in which that task ran.

In Oracle EBS 12.1.1 and 12.2.2, this table is a critical diagnostic resource. Database administrators consult it when a patching run fails, when performance must be analyzed, or when a maintenance session must be audited. The table is populated during the execution of AD utilities and is preserved historically so that prior patch sessions can be reviewed after completion.

From a Data Vault modeling perspective, the mined classification of AD_TASK_TIMING is standalone, which is best interpreted as a satellite-style descriptive table: it holds measurable, volatile attributes (timings, counts, status flags) describing an individual task event, with no enforced hub or link role beyond the single foreign key to JTF_TASK_STATUSES_B. This suggests a design in which AD_TASK_TIMING sits as a descriptive satellite of a task-event concept rather than as a business entity hub itself.

Key Information Stored

The table contains 24 documented columns. The most significant are summarized below. No surrogate primary key column appears explicitly in the documented column list; the closest business-key candidates are SESSION_ID combined with TASK_NUMBER, and WORKER_ID combined with TASK_NUMBER, which uniquely identify a task executed within a given maintenance session.

• SESSION_ID — Identifier for the AD utility session (for example a specific adpatch run); the primary grouping key for reporting.
• JOB_NAME — The name of the job or driver being executed, which is central to identifying failed patches.
• PRODUCT — The EBS product short name (for example, GL or AR) to which the task belongs.
• WORKER_ID — The parallel worker number within the session, allowing per-worker analysis.
• INSTALL_GROUP_NUM — The installation group in which the task was scheduled, used in multi-node patch runs.
• PROGRAM / PHASE / PHASE_NAME — The program name and execution phase context, useful for phase-by-phase performance review.
• ARGUMENTS / COMMAND / SUBDIRECTORY — The actual command line, arguments, and working directory used for the task, aiding reproduction.
• TASK_NUMBER — Sequential ordinal of the task within the session; combined with SESSION_ID, it identifies the individual task.
• START_TIME, END_TIME, ELAPSED_TIME, RESTART_TIME — Timing attributes that support duration and performance analysis.
• RESTART_COUNT / DEFER_COUNT — Counts of how many times the task was restarted or deferred, useful for detecting instability.
• TASK_STATUS_ID — Foreign key to JTF_TASK_STATUSES_B, providing the current status of the task.
• CURRENT_FLAG — Indicates whether the row reflects the current attempt for a given task.
• CREATION_DATE, LAST_UPDATE_DATE, CREATED_BY, LAST_UPDATED_BY — Standard EBS audit columns.

Common Use Cases and Queries

Typical usage focuses on troubleshooting failing patch sessions and on measuring task durations. Reported queries include:

• Listing all tasks in the most recent session: SELECT JOB_NAME, PHASE_NAME, START_TIME, END_TIME, ELAPSED_TIME FROM APPLSYS.AD_TASK_TIMING WHERE SESSION_ID = :sid ORDER BY TASK_NUMBER.
• Finding the longest-running tasks: SELECT JOB_NAME, ELAPSED_TIME FROM APPLSYS.AD_TASK_TIMING WHERE SESSION_ID = :sid ORDER BY ELAPSED_TIME DESC.
• Identifying failed tasks via the status join: SELECT t.JOB_NAME, t.PHASE_NAME, s.NAME FROM APPLSYS.AD_TASK_TIMING t, JTF_TASK_STATUSES_B s WHERE t.TASK_STATUS_ID = s.STATUS_ID AND t.SESSION_ID = :sid.
• Detecting instability through restart/defer counters: SELECT JOB_NAME, RESTART_COUNT, DEFER_COUNT FROM APPLSYS.AD_TASK_TIMING WHERE SESSION_ID = :sid AND (RESTART_COUNT > 0 OR DEFER_COUNT > 0).
• Per-product roll-up reporting for patch performance benchmarking across maintenance windows.

Related Objects

The single documented foreign key relates this table to JTF_TASK_STATUSES_B, joined on TASK_STATUS_ID = STATUS_ID, providing the status name and meaning for each task. Practically, AD_TASK_TIMING is most closely associated with the sibling AD tables that record session-level and phase-level information — the AD_PATCH_RUNS and AD_PATCH_DRIVER tables that identify the patch session referenced by SESSION_ID, the AD_TASK_TIMING-related phase and job listings used by adpatch, and AD_APPLIED_PATCHES for verifying which patches were ultimately applied. The AD utility log files (in the AD_TOP/admin/log directory) mirror the same session and task identifiers stored here and are commonly correlated with these rows. Because session data is written by adpatch and adadmin, the AD utilities themselves are the primary consumers, and DBAs rely on AD_TASK_TIMING as the authoritative post-run record for auditing and diagnosing patch execution.