PCDdebugger Utility for PCD Troubleshooting

PCDdebugger

PCDdebugger is a command-line Python script designed to simplify and accelerate the troubleshooting process for PCD environments. It automates the collection of diagnostic information for various services and resources, consolidating the output into a structured directory for easy analysis and sharing.

Features

• Comprehensive Data Collection: Gathers detailed information for key PCD resources including VMs (Nova), images (Glance), networks (Neutron), ports, volumes (Cinder), stacks (Heat), and users (Keystone).
• Kubernetes Integration: Performs a complete MySQL dump from a specified Kubernetes namespace, essential for debugging control plane issues.
• Dependency Traversal: Automatically discovers and collects data for resources related to a specified VM, such as its ports, volumes, network, subnets, image, and flavor.
• Organized Output: Saves all collected data into a timestamped directory, with subfolders for each service, making the information easy to navigate.
• Archive Option: Includes a --zip flag to automatically create a compressed archive of the collected data, ready for sharing.

Prerequisites

Before running the script, ensure the following are installed and configured on your machine:

• Python 3.6+
• pcd client: Authenticated and configured to connect to your pCD cloud. (Ensure your rc file is sourced).
• kubectl: Authenticated and configured to connect to your Kubernetes cluster.( This is required only in case of mysql dump)
• yq: A command-line YAML processor, used for parsing data from Consul.

Installation

1. Clone the repository:

$ git clone https://github.com/platform9/PCDDebugger.git

2. cd PCDdebugger
3. Make the script executable (optional but recommended):

$ chmod +x ./pcddebugger.py

Usage

The basic command structure is:

./pcddebugger.py [RESOURCE_FLAG] [OPTIONS]

Examples

• Collect all information for a specific VM: This will gather details of the VM, its ports, volumes, network, subnets, image, and flavor.

./pcddebugger.py --vm <VM_UUID>

• Collect details for a specific Glance image:

./pcddebugger.py --image <IMAGE_UUID>

• Collect details for a Neutron network and its subnets: .

./pcddebugger.py --network <NETWORK_UUID>

• Perform a MySQL dump from a Kubernetes cluster: The --namespace flag is required for this operation.

./pcddebugger.py --mysql-dump --namespace <WORKLOAD_REGION>

• Combine multiple flags and create a zip archive:

./pcddebugger.py --vm <VM_UUID> --mysql-dump --namespace <K8S_NAMESPACE> --zip

• Specify a custom output directory and create a zip archive:

./pcddebugger.py --vm <VM_UUID> --output <PATH_TO_DIRECTORY> --zip

./pcddebugger.py --vm aa1bb1cc3-dd4ee5ff6-gg7hh8-ii9jj10kk --output /tmp --zip

Command-Line Flags:

• --vm <ID_OR_NAME>: Collect details for a specific Nova VM and its related resources.
• --image <ID_OR_NAME>: Collect details for a specific Glance image.
• --network <ID_OR_NAME>: Collect details for a specific Neutron network and its subnets.
• --port <ID_OR_NAME>: Collect details for a specific Neutron port.
• --volume <ID_OR_NAME>: Collect details for a specific Cinder volume.
• --stack <ID_OR_NAME>: Collect details for a specific Heat stack.
• --user <ID_OR_NAME>: Collect details for a specific Keystone user.
• --mysql-dump: Perform a MySQL dump of all databases. Requires --namespace.
• --namespace <NAMESPACE>: The Kubernetes namespace to use for the MySQL dump.
• --output <DIRECTORY>: Specify a custom directory for the output files.
• --zip: Compress the final output directory into a .zip file.
• --help: Show the help message and exit.

What It Collects

Status of all Compute, Networking, Image and Storage Services and many more.