To help make switch setup a breeze. CANU can automatically parse CSI output or the Shasta SLS API for switch IPv4 addresses. In order to parse CSI output, use the --sls-file SLS_FILE flag to pass in the folder where the sls_file.json file is located. 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.
If used, the 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 sls_input_file.json file is normally found in the the directory
/var/www/ephemeral/prep/SYSTEMNAME/ - Later in the install process, the 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.
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.txtTo 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| Option | Description |
|---|---|
--sls_file |
File containing the CSI json |
--auth_token |
Token for SLS authentication |
--sls_address |
The address of SLS |
--out |
Name of the output file |
