Collect Linux and setup performance data collection

Edit online

This document provides an example how to collect data for Linux® and the setup for the performance data collection.

About this task

The IBM® technical support team has a specific process in place that helps you to collect Linux data and to setup a performance data collection. The data requested contains dbginfo and sadc data.

Procedure

See and follow the steps in this document:
  • Could you please gather some data for us to look at in this particular situation?
Specifically we would like to request dbginfo and sadc data for understanding the environment better 
during the issue situation.
A description is to be found here: 
https://www.ibm.com/support/knowledgecenter/en/linuxonibm/com.ibm.linux.z.lxsv/lxsv_ts_tool_dbginfo.html

Linux dbginfo and performance Data gathering:
-------------------------------------------------------
1) Go to tmp directory and gather dbginfo
  # cd /tmp
  # /sbin/dbginfo.sh

2) kick off a special sadc run with a 10 sec interval.
  # /usr/lib64/sa/sadc -S XALL 10 /tmp/server_sadc.out &

3) run and finish < the recreation steps / run the test>

4) cancel the running sadc with
  # killall sadc

5) extract human readable sar report from the sadc output file
  # sar -A -t -f /tmp/server_sadc.out > server_sar.txt 
Please give the "server_sar.txt" a meaningful name, e.g. including the linux guest name and date.

6) Gather dbginfo again so we can see deltas for specific counters like interrupt etc.
  # /sbin/dbginfo.sh

7) upload the following files from the /tmp directory to the support ticket (ECUREP). 
  • both dbginfo files from this run
  • server_sadc.out (resp. the name you chose in the sadc command)
  • server_sar.txt (resp. the name you chose in the sar command)