Prepare For Upgrade

Before beginning an upgrade to a new version of CSM, there are a few things to do on the system first.

Reduced resiliency during upgrade

Warning: Management service resiliency is reduced during the upgrade.

Although it is expected that compute nodes and application nodes will continue to provide their services without interruption, it is important to be aware that the degree of management services resiliency is reduced during the upgrade. While one node is being upgraded, if another node of the same type has an unplanned fault that removes it from service, then this may result in a degraded system. For example, if there are three Kubernetes master nodes and one is being upgraded, the quorum is maintained by the remaining two nodes. If one of those two nodes has a fault before the third node completes its upgrade, then quorum would be lost.

Preparation steps

1. Start typescript

  1. (ncn-m001#) If a typescript session is already running in the shell, then first stop it with the exit command.

  2. (ncn-m001#) Start a typescript.

    script -af /root/csm_upgrade.$(date +%Y%m%d_%H%M%S).prepare_for_upgrade.txt
export PS1='\u@\H \D{%Y-%m-%d} \t \w # '

If additional shells are opened during this procedure, then record those with typescripts as well. When resuming a procedure after a break, always be sure that a typescript is running before proceeding.

2. Ensure latest documentation is installed

Before following the steps to prepare for the upgrade, make sure that the latest CSM documentation RPMs are installed on any NCNs where preparation procedures are being performed. These should be for the CURRENT CSM version on the system – not the target version of the upgrade. See Check for latest documentation for instructions.

3. Export Nexus data

Warning: This process can take multiple hours where Nexus is unavailable and should be done during scheduled maintenance periods.

Prior to the upgrade it is recommended that a Nexus export is taken. This is not a required step but highly recommend to protect the data in Nexus. If there is no maintenance period available, then skip this step until after the upgrade process.

Reference Nexus Export and Restore Procedure for details.

4. Adding switch admin password to Vault

If CSM has been installed and Vault is running, add the switch credentials into Vault. Certain tests (for example, goss-switch-bgp-neighbor-aruba-or-mellanox) use these credentials to test the state of the switch. This step is not required to configure the management network. If Vault is unavailable, then this step can be temporarily skipped. Any automated tests that depend on the switch credentials being in Vault will fail until they are added.

  1. (ncn-mw#) Write the switch admin password to the SW_ADMIN_PASSWORD variable if it is not already set.

    read -s SW_ADMIN_PASSWORD

    Note: The use of read -s is a convention used throughout this documentation which allows for the user input of secrets without echoing them to the terminal or saving them in history.

  2. (ncn-mw#) Run the following commands to add the switch admin password to Vault.

    VAULT_PASSWD=$(kubectl -n vault get secrets cray-vault-unseal-keys -o json | jq -r '.data["vault-root"]' |  base64 -d)
alias vault='kubectl -n vault exec -i cray-vault-0 -c vault -- env VAULT_TOKEN="$VAULT_PASSWD" VAULT_ADDR=http://127.0.0.1:8200 VAULT_FORMAT=json vault'
vault kv put secret/net-creds/switch_admin admin=$SW_ADMIN_PASSWORD

5. Ensure SNMP is configured on the management network switches

To ensure proper operation of the River Endpoint Discovery Service (REDS) hardware discovery process, Power Control Service (PCS)/Redfish Translation Service (RTS) management switch availability monitoring, and the Prometheus SNMP Exporter, validate the following:

  • SNMP is enabled on the management network switches.
  • The SNMP credentials on the switches match the credentials stored in all of the following locations:
    • Vault
    • customizations.yaml (stored as a sealed secret)
    • SNMP custom configuration, if applicable. If a custom configuration was used with the CSM Automatic Network Utility (CANU) when generating the management network switch configurations, also check that the credentials in this custom configuration match.

These checks help avoid failure scenarios that can impact the ability to add new hardware to the system. It is not uncommon for CSM upgrades to be paired with system maintenance such as hardware layout changes, expansion, or management network upgrades. If management network switches are reconfigured or new switches are added, and a custom CANU configuration with SNMP settings was not used, it is possible that an administrator may unknowingly push new switch configurations that omit SNMP. If in the process of fixing SNMP, an administrator then adds SNMP credentials to the switches that do not match what is stored in Vault and customizations.yaml, then the resulting REDS, PCS/RTS, and Prometheus errors can be difficult to diagnose and resolve.

CANU custom configuration files should be stored in a version controlled repository so that they can be re-used for future management network maintenance.

For more information, see Configure SNMP. That page contains the following relevant information:

  • Links to vendor-specific switch documentation, which provides more information about configuring SNMP on the management switches.
  • Other SNMP information related to REDS hardware discovery, PCS/RTS management switch availability monitoring, and the Prometheus SNMP Exporter
  • Links to related procedures with Vault, customizations.yaml, sealed secrets, and more.

Return here after verifying that SNMP is properly configured on the management network switches.

6. Running sessions

Boot Orchestration Service (BOS), Configuration Framework Service (CFS), Compute Rolling Upgrade Service (CRUS), Firmware Action Service (FAS), and Node Memory Dump (NMD) sessions should not be started or underway during the CSM upgrade process.

  1. (ncn-m001#) Ensure that these services do not have any sessions in progress.

    This System Admin Toolkit (SAT) command has shutdown as one of the command line options, but it will not start a shutdown process on the system.

    Note: This step may take a longer time if there are many BOS sessions. It is recommended to keep the sessions count minimal to reduce the overall time taken.

    sat bootsys shutdown --stage session-checks

    Example output:

    Checking for active BOS sessions.
Found no active BOS sessions.
Checking for active CFS sessions.
Found no active CFS sessions.
Checking for active FAS actions.
Found no active FAS actions.
Checking for active NMD dumps.
Found no active NMD dumps.
No active sessions exist. It is safe to proceed with the shutdown procedure.

    If active sessions are running, then either wait for them to complete, or shut down, cancel, or delete them.

  2. Coordinate with the site to prevent new sessions from starting in these services.

    There is currently no method to prevent new sessions from being created as long as the service APIs are accessible on the API gateway.

7. Health validation

  1. Validate CSM health.

    Run the CSM health checks to ensure that everything is working properly before the upgrade starts. After the upgrade is completed, another health check is performed, and it is important to know if any problems observed at that time existed prior to the upgrade.

    IMPORTANT: See the CSM Install Validation and Health Checks procedures in the documentation for the CURRENT CSM version on the system. The validation procedures in the CSM documentation are only intended to work with that specific version of CSM.

  2. Validate Lustre health.

    If a Lustre file system is being used, then see the ClusterStor documentation for details on how to validate Lustre health.

8. Stop typescript

For any typescripts that were started during this preparation stage, stop them with the exit command.