Skip to content

Initialization

To help make switch setup a breeze. CANU can automatically parse SLS JSON data - including CSI sls_input_file.json output or the Shasta SLS API for switch IPv4 addresses.

canu init

Initialize CANU by extracting all the switch IPs from CSI generated json, or by getting IPs from SLS.

To access the SLS API, a token must be passed in using the ‘–auth-token’ flag. - Tokens are typically stored in ‘~./config/cray/tokens/’ - Instead of passing in a token file, the environmental variable SLS_TOKEN can be used.

To initialize using JSON instead of the SLS API, pass in the file containing SLS JSON data (normally sls_input_file.json) using the ‘–sls-file’ flag

If used, CSI-generated sls_input_file.json file is generally stored in one of two places depending on how far the system is in the install process. - Early in the install process, when running off of the LiveCD the CSI sls_input_file.json file is normally found in the the directory ‘/var/www/ephemeral/prep/SYSTEMNAME/’ - Later in the install process, the CSI sls_input_file.json file is generally in ‘/mnt/pitdata/prep/SYSTEMNAME/’

The output file for the canu init command is set with the –out FILENAME flag.

noqa: D301, B950

Args:

ctx: CANU context settings
sls_file: File containing the CSI json data
auth_token: Token for SLS authentication
sls_address: The address of SLS
network: Switch network e.g. (CAN, MTL, NMN).
out: Name of the output file

canu init [OPTIONS]

Options

--sls-file( )

File containing system SLS JSON data.

--auth-token( )

Token for SLS authentication

--sls-address( )

  • Default

    api-gw-service-nmn.local

--network( )

Switch network e.g. (CAN, MTL, NMN)

  • Default

    NMN

--out( )

Required Output file with CSI IP addresses

Environment variables

SLS_TOKEN()

Provide a default for --auth-token

Examples

1. CSI

To get the switch IP addresses from CSI output, run the command:

$ canu init --sls-file SLS_FILE --out output.txt
8 IP addresses saved to output.txt

image

2. SLS Shasta API

To get the switch IP addresses from the Shasta SLS API, run the command:

$ canu init --auth-token ~./config/cray/tokens/ --sls-address 1.2.3.4 --out output.txt
8 IP addresses saved to output.txt

CSI Input

  • In order to parse CSI output, use the --sls-file FILE flag to pass in the folder where an SLS JSON file is located.

The CSI sls_input_file.json file is generally stored in one of two places depending on how far the system is in the install process.

  • Early in the install process, when running off of the LiveCD the CSI sls_input_file.json file is normally found in the the directory /var/www/ephemeral/prep/SYSTEMNAME/

  • Later in the install process, the CSI sls_input_file.json file is generally in /mnt/pitdata/prep/SYSTEMNAME/

  • The SLS file can also be obtained from an NCN that’s in the k8s cluster by running cray sls dumpstate list --format json

  • The switch IPs will be read from the ‘NMN’ network, if a different network is desired, use the --network flag to choose a different one e.g. (CAN, MTL, NMN).

To get the switch IP addresses from CSI output, run the command:

canu init --sls-file SLS_FILE --out output.txt

Potential output:

8 IP addresses saved to output.txt

SLS API Input

  • To parse the Shasta SLS API for IP addresses, ensure that you have a valid token. The token file can either be passed in with the --auth-token TOKEN_FILE flag, or it can be automatically read if the environmental variable SLS_TOKEN is set. The SLS address is default set to api-gw-service-nmn.local, if you are operating on a system with a different address, you can set it with the --sls-address SLS_ADDRESS flag.

To get the switch IP addresses from the Shasta SLS API, run the command:

canu init --auth-token ~./config/cray/tokens/ --sls-address 1.2.3.4 --out output.txt

Potential output:

8 IP addresses saved to output.txt

image

The output file for the canu init command is set with the --out FILENAME flag.