Rollback XCO

Initiate a rollback when there is a deployment failure.

About this task

Follow this procedure to rollback a deployment failure.

Procedure

  1. Unwind the partial installation or undeploy the failed XCO instance.
    no efa deploy
  2. Copy the XCO instance.
    efa deploy
  3. Use system backups available in the /apps/efa_logs/backup/ directory or copy the required backup files to the /apps/efa_logs/backup/ directory.
    $ scp root@10.20.48.170:/home/user/EFA-3.4.0-GA-2023-08-20T07-08-43.921.tar /apps/efa_logs/backup/
    The authenticity of host '10.20.48.170 (10.20.48.170)' can't be established.
    ECDSA key fingerprint is SHA256:rQYa5NjeFWtLvCCUzjELs+9jd/6E+hBeEeHIYdFBs2I.
    Are you sure you want to continue connecting (yes/no)? yes
    Warning: Permanently added '10.20.48.170' (ECDSA) to the list of known hosts.
    root@10.20.48.170's password:
    EFA-3.4.0-GA-2023-08-20T07-08-43.921.tar                                      100%  732KB  18.6MB/s   00:00
  4. Log in to the XCO system as an Extreme user.
  5. Restore the XCO configuration.
    efa system restore -–backup-tar <file_name>
  6. Post XCO rollback, to verify whether there are any devices in the cfg refresh error state, run the following command:
    efa fabric show
    • To remove the devices from the cfg refresh error state, run the following command:
      efa inventory device update --ip <device_ip>