ClusterInformationAggregationTool (CIAT)

This page is for a staff meeting presentation. Usage instructions can be found on the project repository.

Problem

• ~140 nodes of the tembo cluster need to be added to inventory
  • Nodes are not all identical, minor subtleties
• Helpful if barcode were written into BMC

Tech Stack

• Python chosen as scripting language
  • I am familiar with it
  • High level features available ex. JSON Parsing
• Salt Stack
  • Used to run script to collect information from all nodes (Salt selector "*")
  • Can be optionally used to associate barcodes with hostnames
• lshw, lscpu
  • Simple tools to get system specs with JSON output
  • Available though most package managers
• Input and output as CSV files
  • Easy to work with in python
  • Compatible with CSCF inventory
• IPMI tool
  • Only option for writing barcodes into BMC FRU
  • In the form of a binary packaged with CIAT

Usage Overview

Full details in repo Readme

Download into a shared directory accessible by all nodes:

git clone https://git.uwaterloo.ca/qhyun/CIAT.git

Fill out example_data.csv and make accessible to the head node. Give the command:

./ciat.py --output /share/output master /share/example_data.csv

You get out:

tem99.tembo-domain.cs.uwaterloo.ca:
    2019-10-18 13:08:49 INFO     Running as slave
    2019-10-18 13:08:49 INFO     Running lshw
    2019-10-18 13:08:49 INFO     Verifying lshw exists
    2019-10-18 13:09:00 INFO     lshw installed, running now
    2019-10-18 13:09:01 INFO     Determining simple lshw fields
    2019-10-18 13:09:01 INFO     Determining total amount of ram
    2019-10-18 13:09:01 INFO     Determining network info
    2019-10-18 13:09:01 WARNING  Could not load ip of network eno2 / MAC 0c:c4:7a:d9:0c:e9. This network is most likely useless / internal.
    2019-10-18 13:09:01 INFO     Network IDs    : eno1, enp4s0
    2019-10-18 13:09:01 INFO     Network MACs   : 0c:c4:7a:d9:0c:e8, 24:8a:07:79:7f:a0
    2019-10-18 13:09:01 INFO     Network IPs    : 192.168.152.199, 192.168.252.199
    2019-10-18 13:09:01 INFO     Network Speeds : 1Gbit/s, 10Gbit/s
    2019-10-18 13:09:01 INFO     lshw completed and parsed
    2019-10-18 13:09:01 INFO     Running lscpu
    2019-10-18 13:09:01 WARNING  Could not load lscpu as JSON data. Please upgrade lscpu to a version which supports --json. Now trying to fetch data though regex
    2019-10-18 13:09:01 INFO     lscpu completed and parsed
    2019-10-18 13:09:01 INFO     Running salt stack to determine barcode
    2019-10-18 13:09:03 INFO     Running IPMICFG to burn barcode CS010574
    2019-10-18 13:09:04 INFO     Successfully wrote data to file /share/tools/invadd/./output/tem99
    2019-10-18 13:09:04 INFO     Information collection complete
tem54.tembo-domain.cs.uwaterloo.ca:
    Minion did not return. [Not connected]
tem74.tembo-domain.cs.uwaterloo.ca:
    Minion did not return. [No response]

You can find the output file at ./your_output_path/aggregate.csv

If you see this I forgot to put the file here

CSCF Inventory should now happily consume it.

Links

Git Repo

RT