Dell PowerFlex 4.5.x Administration Guide

Gracefully shut down or reboot a node

Prepare the node for a patching or maintenance operation (such as a part replacement) by entering the node into maintenance mode and shutting down/rebooting the node in a graceful fashion.

1. When shutting down/rebooting a node that is a primary MDM (manager), it is recommended that you manually switch MDM ownership to a different node:
   1. From the PowerFlex CLI (SCLI), run:

      scli --query_cluster

   2. If the node's IP addresses are included in the --query_cluster output, the faulty node has a role of either MDM or tiebreaker, in addition to its SDS role.
      If the node's IP address is located in the primary MDM role, a switch-over action is required.
   3. Switch MDM ownership to a different node:

      scli -switch_mdm_ownership (-new_primary_mdm_id <ID> | --new_primary_mdm_ip <IP> | --new_primary_mdm_name <NAME>)

      The node remains in the cluster. The cluster will be in degraded mode after it is powered off, until the faulty component or patch operation in the node is fixed and the node is powered back on.
   4. Verify that the cluster status shows that the node is not the primary MDM anymore:

      scli --query_cluster

      Output similar to the following should appear, with the relevant node configuration and IP addresses for your deployment:

      Cluster:
          Mode: 5_node, State: Normal, Active: 5/5, Replicas: 3/3
          Virtual IP Addresses: 9.20.10.100, 9.20.110.100
      Primary MDM:
          ID: 0x775afb2a65ef1f02
              IP Addresses: 9.20.10.104, 9.20.110.104, Management IP Addresses: 10.136.215.239, Port: 9011, Virtual IP interfaces: sio_d_1, sio_d_2
              Version: 2.0.13000
      Secondary MDMs:
          ID: 0x5b2e9f273b7af9b0
              IP Addresses: 9.20.10.105, 9.20.110.105, Management IP Addresses: 10.136.215.223, Port: 9011, Virtual IP interfaces: sio_d_1, sio_d_2
              Status: Normal, Version: 2.0.13000
          ID: 0x5828f65b15e778f1
              IP Addresses: 9.20.10.102, 9.20.110.102, Management IP Addresses: 10.136.215.232, Port: 9011, Virtual IP interfaces: sio_d_1, sio_d_2
              Status: Normal, Version: 2.0.13000
      Tiebreakers:
          ID: 0x6618e0b804644ca4
              IP Addresses: 9.20.10.101, 9.20.110.101, Port: 9011
              Status: Normal, Version: 2.0.13000
          ID: 0x12534ccb3d28fee3
              IP Addresses: 9.20.10.103, 9.20.110.103, Port: 9011
              Status: Normal, Version: 2.0.13000
      

      In the example output, the primary MDM IP addresses are:

      IP Addresses: 9.20.10.104, 9.20.110.104, Management IP Addresses: 10.136.215.239

      The secondary IP addresses are:

      IP Addresses: 9.20.10.105, 9.20.110.105, Management IP Addresses: 10.136.215.223
      IP Addresses: 9.20.10.102, 9.20.110.102, Management IP Addresses: 10.136.215.232

2. Move all applications to a different node:
   • On an VMware ESXi node that is not a cluster member, and that is not configured for HA and DRS, migrate the VMs to another VMware ESXi.
   • On a Linux node, migrate the applications (or the VMs, if the node is running a hypervisor).
3. Log in to PowerFlex Manager as an admin user.
4. On the menu bar, click Block > SDSs.
5. Select the relevant SDS and click More Actions > Enter Maintenance Mode.
6. In the Enter SDS <name> into Maintenance Mode dialog box, select one of the following options:
   • Instant
   • Protected
7. Click Enter Maintenance Mode.
8. In the confirmation message dialog box, click Dismiss.
9. If you are applying a patch:
   1. Run the patch.
   2. Reboot the node, if necessary.