Scheduler Errand Fails at Dynatrace OneAgent Pre-Start Script

Products

• VMware Tanzu Platform - Cloud Foundry

Issue/Introduction

During patching or deployment activities, the deploy-scheduler errand fails. The BOSH task output contains an error similar to: Action Failed get_task: Task [ID] result: 1 of 3 pre-start scripts failed. Failed Jobs: dynatrace-oneagent
Environment
VMware Tanzu Platform - Cloud Foundry
Scheduler Tile installed with Dynatrace OneAgent add-on/job.

Cause

The dynatrace-oneagent pre-start script fails to execute or complete because the VM assigned to the scheduler instance group does not have enough persistent or ephemeral disk space to accommodate the OneAgent requirements during initialization.

Resolution

To resolve this issue, increase the disk allocation for the scheduler instances:
1. Identify the Failure: Run the errand with the keep-alive flag to prevent the VM from being deleted upon failure: bosh -d [deployment-name] run-errand deploy-scheduler --keep-alive

2. Verify Disk Constraints:

3. SSH into the errand VM.

4. Navigate to /var/vcap/sys/log/dynatrace-oneagent/ and review pre-start.stdout.log or pre-start.stderr.log.

5. Check current disk usage using df -h.

6. Increase Resource Allocation:
    Log into Ops Manager.
    Click on the Scheduler tile.
    Select Resource Config.
    Locate the Scheduler (or the relevant errand) job and increase the Persistent Disk or Ephemeral Disk size.
    Click Save.

7. Apply Changes:
    Return to the Ops Manager Installation Dashboard.
    Click Review Pending Changes.
    Click Apply Changes to redeploy the scheduler with the updated disk configuration.