Articles in this section

Replacing a CNode

Sara Levy
  • Updated
Follow

Caution

When deactivating CNodes, make sure to maintain the following ratios to maintain cluster viability and data availability:

  • For larger clusters, at least 1 fully operational CNode must be available for every 20 operational DNodes.

  • For smaller clusters, at least 2 fully operational CNodes must be available.

Perform this procedure if you need to replace a CNode. This requires powering off the CNode and removing it from the CBox chassis.

Note

To request a field replacement unit (FRU) for a faulty component, contact VAST Data Support.

This replacement procedure uses the power off and replace features in VMS.

When the procedure is complete:

  • Passwords that were configured on the removed CNode are configured on the replacement CNode.

  • Network settings are configured on the replacement CNode to match the network configuration of the removed CNode.

  • SSH keys are updated on the replacement CNode

  • The VAST OS is updated to the correct version.

Replace the CNode

  1. Power off the CNode using one of the following methods:

    • From the VAST Web UI:

      1. From the left menu, select Infrastructure and then CNodes.

      2. On the CNodes list, identify the CNode you want to replace, click Actions_Button.png to open the Actions menu and select Power off.

    • From the VAST CLI:

      1. Identify the ID of the CNode you want to replace:

        vcli: admin> cnode list
             +----+--------------+--------------+---------------------------+-------------+-----------+--------+---------+----------------------------------------+------------+------------+------------+-------------+-------+-------------
             | ID | Name         | IP           | IPv6                      | Mgmt-ip     | Host Name | State  | Enabled | Build                                  | Management | Cluster    | OS-Version | CBox        | Cores | Serial
             +----+--------------+--------------+---------------------------+-------------+-----------+--------+---------+----------------------------------------+------------+------------+------------+-------------+-------+-------------
             | 1  | cnode-128-21 | 10.10.128.21 | fe80::9a03:9bff:fe8f:ba4c | 10.55.21.11 | c-128-21  | ACTIVE | True    | build-as-orion-96399-835783 | Yes        | vast42-kfs | 12.8.16    | cbox-501023 | 14    | a4:bf:01:5c:
             | 2  | cnode-128-22 | 10.10.128.22 | fe80::a6bf:1ff:fe5c:eaa4  | 10.55.21.13 | c-128-22  | ACTIVE | True    | build-as-orion-96399-835783 | No         | vast42-kfs | 12.9.1     | cbox-501023 | 14    | a4:bf:01:5c:
             | 3  | cnode-128-23 | 10.10.128.23 | fe80::9a03:9bff:fe8f:ba44 | 10.55.21.15 | c-128-23  | ACTIVE | True    | build-as-orion-96399-835783 | No         | vast42-kfs | 12.8.16    | cbox-501023 | 14    | a4:bf:01:5c:
             +----+--------------+--------------+---------------------------+-------------+-----------+--------+---------+----------------------------------------+------------+------------+------------+-------------+-------+-------------

      2. Run the following command with the relevant ID. For example, if you are replacing the failed CNode above:

        vcli: admin> CNode deactivate --id 2 --power-off

    The power off option deactivates and powers off the CNode. The CNode state transitions to DEACTIVATING and then to INACTIVE.

  2. Monitor the modify_CNode activity:

    • From the VAST Web UI: Go the Activities page and check the state of the modify_CNode activity.

    • From the VAST CLI: run vtask list and check the state of the modify_CNode activity.

  3. When the state of the modify_CNode activity is COMPLETED, disconnect all QSFP28 and RJ-45 cables from the CNode.

    Label the cables if needed so that you will be able to reconnect the cables to the same ports on the replacement CNode after replacement. e.g. If the QSFP28 cables are not labeled, label the left and right cables before disconnecting.

  4. Remove the CNode.

    To remove a CNode from an Ice Lake (HPE) model CBox:

    1. Disconnect all peripheral cables from the server.

    2. Loosen the release lever thumbscrew (1) by turning it counter clockwise.

    3. Pull the release lever to pull the server out of the bay (2).

      1-80.jpg

  5. Insert the replacement CNode.

    For an Ice Lake (HPE) model CNode:

    1. Slide in the replacement CNode fully and push the release lever (2) into the locked position.

    2. Tighten the release lever thumbscrew (1) by turning it clockwise until it is hand tight.

  6. Reconnect the RJ45 and QSFP28 cables to the equivalent ports to which they were previously connected.

  7. When the cables are connected and the CNode state is ACTIVE, do one of the following:

    • From the VAST Web UI:

      1. From the left menu, select Infrastructure and then CNodes.

      2. On the CNodes list, identify the same CNode entry again, click Actions_Button.png to open the Actions menu and select Replace.

    • From the VAST CLI:

      • Run the following command with the relevant ID:

        vcli: admin> CNode activate --id ID --replace

        Note

        If any other CNode or CNode in the cluster is known to be in a non-active state, or if one of a CNode's or CNode's data network ports is disconnected from the switch, then add the additional parameter --skip-network-validation.

        That is, run this command:

        vcli: admin> CNode activate --id ID --replace --skip-network-validation

    The replace process now updates the passwords, serial number, network configuration, SSH keys, and OS version.

    The CNode begins to power on automatically. It transitions from INACTIVE to ACTIVATING and then to ACTIVE.

  8. Monitor the modify_CNode activity again.

    When the replacement is complete, the modify_CNode activity is reported as COMPLETED.

Related articles

Comments

0 comments

Article is closed for comments.