Back to all use cases

AssessmentCross-Industry

A written list of what could fail the next batch window, ranked before anything fails

Data Platform Team

A typical engagement for a data team that runs Informatica PowerCenter nightly and has never had the repository, the Integration Service settings, the session logs and the schedules reviewed together: a ranked list of findings, with nothing changed.

Informatica support

Overview

This is a typical engagement for a data platform team that runs a PowerCenter domain to feed a warehouse every night. The people who set the environment up have moved on, the workflows still run, and nobody can say how long the session logs are kept, which sessions would recover cleanly after a failure, or which schedule depends on which. AceMQ engineers and a partner data-integration specialist review the environment read-only.

Challenge

PowerCenter settings that matter in an incident are spread across the Repository Service, the Integration Service, each session and the operating system. A Session task defaults to fail task and continue workflow, so a team that assumes recovery works may find it does nothing. A Failed workflow cannot be recovered at all, according to Informatica's documentation, and log retention that was never set can fill a node's disk. These facts only show up when someone reads the whole setup.

Environment

A PowerCenter domain with one or two nodes, a Repository Service on a database the team runs, one or more Integration Services, several hundred workflows scheduled partly in PowerCenter and partly in an external scheduler that calls pmcmd, and session logs and recovery files on local disk.

Approach

AceMQ works read-only with the team's own accounts and changes nothing in the environment. The engineers list the services and their settings, read the failed-workflow history, check the retention and recovery settings, and map the schedules and the dependencies between them. Each finding is ranked by what it would cost in a batch window.

Solution

  1. Repository Service review: operating mode, database pool size, connection timeout and whether a repository backup has been taken and restored in a test
  2. Integration Service review: the limit on concurrent Session and Command tasks, resilience timeout, node assignment and what happens to workflows when a node stops
  3. Session recovery review: the recovery strategy on each Session task, which sessions are set to resume from the last checkpoint and whether their sources return repeatable data
  4. Log and disk review: session log retention, whether logs are kept by run count or timestamp, and the size and age of recovery state in $PMStorageDir
  5. Failed-workflow history: which workflows fail most, which fail on the same weekday, and how each was brought back
  6. Schedule and dependency map: what runs in PowerCenter, what runs in the external scheduler, and the order that is held only in people's heads

Outcome

The engagement delivers a written report with every finding ranked by risk, the settings that should change and why, a map of schedules and dependencies, and a runbook outline for the recovery decisions the team makes most often. The team decides what to change; the assessment itself changes nothing.

Technologies

Informatica PowerCenterRepository ServiceIntegration Servicepmcmdpmrep

Ready for an Informatica PowerCenter Health Check?

AceMQ's senior Informatica PowerCenter engineers have handled this exact type of engagement before. Whether you need architectural guidance, hands-on remediation, or an ongoing managed partnership, we're ready to help.