Forget Storage Node

Forget option makes the data store forget about the drive. Use the forget option when you need to remove drives for maintenance or replacement in the following cases:

  • When a node disappears permanently, for example, due to disk removal or failure it is necessary to ‘forget’ it.
  • When you need to remove and repair the membership in order to resync a drive.

Forget command removes the failed storage node from all formerly owned vDisks. Use the forget disk option only if you are going to remove the vDisk completely!


Forget storage node via UI



To forget storage node:

  1. Go to your Control Panel > Storage Nodes menu.
  2. On the screen that appears, you will see the list of all storage nodes in the cloud.
  3. Click the Actions button next to the required storage node, then choose Forget.

To forget several storage nodes:
  1. Go to your Control Panel > Storage Nodes menu.
  2. On the screen that appears, you will see the list of all storage nodes in the cloud.
  3. Tick the check-boxes next to the required storage nodes, then click the Forget Nodes button.
To be able to use the forgotten node again, you need to format it using the external utility or by running the following command via CLI:
formatandconfigure



Forget storage node via CLI



To forget storage node via CLI:

  1. Check if that there are sufficient replicas and good paths for all vDisks that reside on that disk before removing a physical drive. Then:

    • To forget a node from a specified VDisk:

      onappstore forget forgetlist=<UUID> vdisk_uuid=<VDISK_UUID
    • To forget a node member from all VDisks:

      onappstore forgetfromall forgetlist=<UUID>
      
  2. Remove the disk.
  3. After that, you can add a new vDisk at this point.
    Before adding a new disk:

    • Make sure that all vDisks that are hosted on the compute resource have sufficient replicas.
    • Migrate all VSs that are hosted on the compute resource to another compute resource.

  4. For OnApp 6.0
    Skip this step if you are running OnApp 6.0.
    For OnApp 6.1
    To register the disk,you need to deselect the old disk and select the new one via UI. Update the configuration by running the SAN controller update config command:

     

    [root@<ipaddr_hv_1> ~]# SANController updateconfig
  5. After that, add the disk to the data store by editing the Integrated Storage data store settings in the OnApp Control Panel.
  6. Repair the degraded vDisks of the removed member.
  7. Run the following commands in the CLI:

    onappstore repairmembership <UUID>
    onappstore repair <UUID>

     

  8. Check the disk status as before with the repair operation:

    onappstore resynchstatus uuid=mq98y0trhi5bxo 
    status={u'3335881780':{u'4142566975':59},u'138773005':{u'433710490':66}}
    result=SUCCESS


Failed disk removal



If you have a failed disk, you can remove it. To remove a failed disk:

  1. Check for a copy of the content.
  2. Forget disk using forgetlist command.
  3. Repair membership for all vdisks that had content on that disk 
  4. Repair VDisk for all vdisks that had content on that disk 
Steps 3-4 are performed for a vdisk with the repair option from UI.