HMC Enhanced View: Collecting PEDBG from the HMC

Resolving The Problem

This document describes how to collect HMC diagnostic data (pedbg data) for use with problem determination. For issues involving servers managed by two HMCs, the data should be collected on both HMCs.

Section 1. Preparing to Collect the Data

Step 1: (Optional) Enable remote command execution
To collect pedbg remotely, you must:

For further information see HMC Enhanced: Enabling ssh/remote command execution


Step 2: Ensure a user ID of hscpe with a task role of hmcpe exists on the HMC.
The hscpe ID is not a built-in ID like hscroot, it must be created. To administer HMC users, you must be logged in as hscroot or a hmcsuperadmin user.

Locally (at the HMC): Open a restricted shell locally on the HMC by selecting HMC Management in the left frame. Select Console Management > Open a Restricted Shell terminal in the right frame.

Remotely: Open an ssh session (PuTTY) to the HMC.

Section 2. Collecting Data Using PEDBG

Step 1: Log in as user hscpe.
You must be logged in to a restricted shell terminal session as user hscpe with role hmcpe to collect pedbg.

Remotely: Start an ssh (PuTTY) session with user hscpe.

Locally: Log in to the local HMC console as user hscpe. Then open a Restricted Shell terminal windows by selecting HMC Management in the navigation area, then select Console Management > Open a Restricted Shell terminal in the work frame (right frame).
To switch to user ID hscpe when the user is already logged on locally with another user ID run the following command in the restricted shell terminal (Remote Command Execution must be enabled):
ssh hscpe@localhost
If prompted to verify authenticity of the host, reply yes and press enter.  Enter the hscpe password when prompted.

Step 2: Re-create the problem.
If possible, re-create the problem before collecting a pedbg. Otherwise, note the HMC date and time. Notify your support representative of the date and time of the problem being investigated relative to the HMC clock.  If the HMC clock is not current, inform your support representative of this also. For example, "The problem occurred approximately 2:00 PM on March 16. The HMC time is 30 minutes slow so that would be about 1:30 PM HMC time.".

Step 3: Collect pedbg
Note: For problems involving managed servers managed by two HMCs, collect pedbg files from both HMCs.

Run PEDBG command
Run the pedbg command with the options specified by the support representative.

If the support representative did not specify the options (or for issues that need more than a few hours of historic data) type the following:
pedbg -c -q 4

If the problem was re-created immediately before pedbg collection, the "-q 3" option can typically be used, which results in a much smaller file:
pedbg -c -q 3

If the support representative requested dumps from the managed system(s) and HMC data type the following to collect all dumps + all historic data:
pedbg -c -q d4

Section 3. Getting the Data to IBM
There are many ways the data can be sent to IBM. The following is a list of preferred methods, starting with the quickest and easiest method. The method used varies depending on the type of network access available to the HMC and the availability of media.