Dell ObjectScale 1.3 Administration Guide

Reactive Disk Removal Service Procedure (for Appliance and Software Bundle)

1. CSI marks the status of failed disks from IN_USE to RELEASING.
2. The cluster operation CR gets created.

   admin@lehi-dirt:~> kubectl get co -A
   NAMESPACE     NAME                                                                         TYPE               STATUS             OBJECT NAME                                                AGE
   objectscale   diskremoval-cfb0478d-cacb-4c3c-b612-659701949847   DiskRemoval   TriggerRemoval   69d56273-b15b-49d5-bd65-b75b525b3425   19s
   

3. The object store status moves to "ReplacingPV" .

   admin@lehi-dirt:~> kubectl get ecs -A
   NAMESPACE     NAME                 PHASE              READY COMPONENTS   S3 ENDPOINT         MGMT API
   objectscale          dirt-objectstore   ReplacingPV      21/21                                 10.249.248.25:443   10.249.248.27:4443

   NOTE:This change is applicable only for SS pods. For non-SS pods, the object store status does not change.
4. The recovery of the disk starts.

   admin@lehi-dirt:~> kubectl get serviceprocedure -A
   NAMESPACE     NAME                                                                       TYPE         STATUS       OBJECT NAME                                             AGE
   objectscale         recovery-dirt-objectstore-nvmeengine-1-541fc74e   Recovery   Recovering    objectscale/dirt-objectstore-nvmeengine-1   17m
   

5. Once the recovery is completed, the status changes to "ReadytoEject".

   admin@lehi-dirt:~> kubectl get co -A
   NAMESPACE     NAME                                                                               TYPE               STATUS            OBJECT NAME                                              AGE
   objectscale         diskremoval-cfb0478d-cacb-4c3c-b612-659701949847   DiskRemoval   ReadyToEject   69d56273-b15b-49d5-bd65-b75b525b3425   128m
   

6. When the status turns "ReadytoEject", clickEject to blink the disk on the rack.
   A dialogue box is displayed to confirm the eject disk procedure.
7. Click Eject disk.
   The disk is ready to be physically removed.
8. Remove the disk physically.
   You can physically locate the correct disk in two ways:

   1. By checking the CSI-01 alert; see the Monitoring Events, Audits, and Alerts section for more details.
   2. By using the following commands:

   First define the drive name driveName=<drive name> , and then get the node and slot information using the below commands:

   #Node
   nodeUUID=$(kubectl get drive ${driveName} -o yaml | grep NodeId | awk -F ': ' '{print$2}')
   kubectl get csibmnode | grep $nodeUUID | awk '{print $3}'

   #Slot
   kubectl get drive ${driveName} -o yaml | grep Slot | awk -F ': ' '{print$2}'

   Usage of the disk is changed to "Removed", and the procedure is complete.

   admin@lehi-dirt:~> kubectl get co -A
   NAMESPACE     NAME                                                                               TYPE               STATUS    OBJECT NAME                                              AGE
   objectscale         diskremoval-cfb0478d-cacb-4c3c-b612-659701949847   DiskRemoval   Success     69d56273-b15b-49d5-bd65-b75b525b3425   138m
   

   admin@lehi-dirt:~> kubectl get drive 69d56273-b15b-49d5-bd65-b75b525b3425 -o yaml
   apiVersion: csi-baremetal.dell.com/v1
   kind: Drive
   metadata:
     annotations:
       health: bad
       removal: ready
     creationTimestamp: "2023-07-24T13:14:34Z"
     generation: 10
   labels:
   .....
     name: 69d56273-b15b-49d5-bd65-b75b525b3425
   spec:
     Firmware: 1.1.1
     Health: BAD
     IsClean: true
   Status: ONLINE
     Type: NVME
     UUID: 69d56273-b15b-49d5-bd65-b75b525b3425
     Usage: REMOVED
     VID: "0x1179"