Skip to main content
ONTAP MetroCluster

Replacing a shelf nondisruptively in a fabric-attached MetroCluster configuration

Contributors netapp-pcarriga netapp-thomi netapp-martyh

You might need to know how to replace a shelf nondisruptively in a fabric-attached MetroCluster configuration.

Note This procedure is only for use in a fabric-attached MetroCluster configuration.

Disabling access to the shelf

You must disable access to the shelf before you replace the shelf modules.

Check the overall health of the configuration. If the system does not appear healthy, address the issue first before proceeding.

Steps
  1. From both clusters, offline all plexes with disks on the affected shelf stack:

    aggr offline plex_name

    The example shows the commands for offlining plexes for a controller running clustered OTNAP.

    cluster_A_1::> storage aggregate plex offline -aggr aggrA_1_0 -plex plex0
    cluster_A_1::> storage aggregate plex offline -aggr dataA_1_data -plex plex0
    cluster_A_2::> storage aggregate plex offline -aggr aggrA_2_0 -plex plex0
    cluster_A_2::> storage aggregate plex offline -aggr dataA_2_data -plex plex0
  2. Verify that the plexes are offline:

    aggr status -raggr_name

    The example shows the commands for verifying that the aggregates are offline for a controller running cMode.

    Cluster_A_1::> storage aggregate show -aggr aggrA_1_0
    Cluster_A_1::> storage aggregate show -aggr dataA_1_data
    Cluster_A_2::> storage aggregate show -aggr aggrA_2_0
    Cluster_A_2::> storage aggregate show -aggr dataA_2_data
  3. Disable the SAS ports or switch ports depending on whether the bridges connecting the target shelf are connecting a single SAS stack or two or more SAS stacks:

    • If the bridges are connecting a single SAS stack, disable the switch ports that the bridges are connected to using the appropriate command for your switch.

      The following example shows a pair of bridges that connect a single SAS stack, which contains the target shelf:

      mcc shelf replacement bridges with a single stack

      Switch ports 8 and 9 on each switch connect the bridges to the network.

      The following example shows ports 8 and 9 being disabled on a Brocade switch.

      FC_switch_A_1:admin> portDisable 8
      FC_switch_A_1:admin> portDisable 9
      
      FC_switch_A_2:admin> portDisable 8
      FC_switch_A_2:admin> portDisable 9

      The following example shows port 8 and 9 being disabled on a Cisco switch.

      FC_switch_A_1# conf t
      FC_switch_A_1(config)# int fc1/8
      FC_switch_A_1(config)# shut
      FC_switch_A_1(config)# int fc1/9
      FC_switch_A_1(config)# shut
      FC_switch_A_1(config)# end
      
      FC_switch_A_2# conf t
      FC_switch_A_2(config)# int fc1/8
      FC_switch_A_2(config)# shut
      FC_switch_A_2(config)# int fc1/9
      FC_switch_A_2(config)# shut
      FC_switch_A_2(config)# end
    • If the bridges are connecting two or more SAS stacks, disable the SAS ports connecting the bridges to the target shelf:
      SASportDisable port number

      The following example shows a pair of bridges that connect four SAS stacks. SAS stack 2 contains the target shelf:

      mcc shelf replacement bridges with four stacks

      SAS port B connects the bridges to the target shelf. By disabling only SAS port B on both shelves, the other SAS stacks can continue to serve data during the replacement procedure.

      In this case, disable the SAS port connecting the bridge to the target shelf:

      SASportDisable port number

      The following example shows SAS port B being disabled from the bridge and also verifies that it is disabled. You must repeat the command on both bridges.

      Ready. *
      SASPortDisable B
      
      SAS Port B has been disabled.
  4. If you previously disabled the switch ports, verify that they are disabled:

    switchShow

    The example shows that the switch ports are disabled on a Brocade switch.

    FC_switch_A_1:admin> switchShow
    FC_switch_A_2:admin> switchShow

    The example shows that the switch ports are disabled on a Cisco switch.

    FC_switch_A_1# show interface fc1/6
    FC_switch_A_2# show interface fc1/6
  5. Wait for ONTAP to realize that the disk is missing.

  6. Power off the shelf that you want to replace.

Replacing the shelf

You must physically remove all of the cables and the shelf before inserting and cabling the new shelf and shelf modules.

Steps
  1. Remove all disks and disconnect all cables from the shelf that is being replaced.

  2. Remove the shelf modules.

  3. Insert the new shelf.

  4. Insert the new disks into the new shelf.

  5. Insert the shelf modules.

  6. Cable the shelf (SAS or Power).

  7. Power on the shelf.

Reenabling access and verifying the operation

After the shelf has been replaced, you need to reenable access and verify that the new shelf is operating correctly.

Steps
  1. Verify that the shelf powers properly and the links on the IOM modules are present.

  2. Enable the switch ports or SAS port according to the following scenarios:

    Option

    Step

    If you previously disabled switch ports

    1. Enable the switch ports:

      portEnable port number

      The example shows the switch port being enabled on a Brocade switch.

      Switch_A_1:admin> portEnable 6
      Switch_A_2:admin> portEnable 6

      The example shows the switch port being enabled on a Cisco switch.

      Switch_A_1# conf t
      Switch_A_1(config)# int fc1/6
      Switch_A_1(config)# no shut
      Switch_A_1(config)# end
      
      Switch_A_2# conf t
      Switch_A_2(config)# int fc1/6
      Switch_A_2(config)# no shut
      Switch_A_2(config)# end

    If you previously disabled a SAS port

    1. Enable the SAS port connecting the stack to the shelf location:

      SASportEnable port number

      The example shows SAS port A being enabled from the bridge and also verifies that it is enabled.

      Ready. *
      SASPortEnable A
      
      SAS Port A has been enabled.
  3. If you previously disabled the switch ports, verify that they are enabled and online and that and all devices are logged in correctly:

    switchShow

    The example shows the switchShow command for verifying that a Brocade switch is online.

    Switch_A_1:admin> SwitchShow
    Switch_A_2:admin> SwitchShow

    The example shows the switchShow command for verifying that a Cisco switch is online.

    Switch_A_1# show interface fc1/6
    Switch_A_2# show interface fc1/6
    Note After several minutes, ONTAP detects that new disks have been inserted and displays a message for each new disk.
  4. Verify that the disks have been detected by ONTAP:

    sysconfig -a

  5. Online the plexes that were offline earlier:

    aggr onlineplex_name

    The example shows the commands for placing plexes on a controller running cMode back online.

    Cluster_A_1::> storage aggregate plex online -aggr aggr1 -plex plex2
    Cluster_A_1::> storage aggregate plex online -aggr aggr2 -plex plex6
    Cluster_A_1::> storage aggregate plex online -aggr aggr3 -plex plex1

    The plexes begin to resynchronize.

    Note You can monitor the progress of resynchronization using the aggr status -raggr_name command.