The replacement of the isilon disk is not directly plugged and unplugged, and some commands are required to operate. If you are unsure, add VX: StorageExpert
Note: There are three types of commands,
Step 1: first ssh or serial port to the node to replace the hard disk
Step 2: isi devices drive list after OneFS 8.0
OneFS 8.0 and earlier isi devices
The location of this disk is SMARTFAIL to be problematic.
If the status of the disk is smartfail, the disk cannot be replaced directly. Smartfail means that the disk is being reconstructed. Only when the disk status is REPLACE or empty can the following command be executed for replacement.
Step 3: If the disk to be replaced has already smartfailed, you do not need to perform the smartfail action on the disk, otherwise perform the following smartfail action
- OneFS 8.0 以后 isi devices drive smartfail <bay-number> --node-lnn <node-number>
- OneFS 8.0 以前 isi devices -a smartfail -d <node-number>:bay<bay-number>
X200-1# isi devices -a smartfail -d 1:9
Error connecting to isi_rpc_d on node id 1: timed out
!! Error: the smartfail action is invalid for a missing drive.
Step 4: Run step 2 again to confirm that the status of the hard disk to be replaced is REPLACE/EMPTY (if the node has been restarted, it will become empty)
Step 5: Physically remove the faulty hard drive and insert a new hard drive
Step six:
Then check the status of preparing
- After OneFS 8.1, OneFS will automatically add hard drives. If ARR (Automatic Replacement Recognition) is disabled, you need to use the following 8.0 command to increase the drive
- OneFS 8.0 isi devices drive add <bay-number) --node-lnn <node-number>
Add after formatting
- OneFS 8.0 以前 isi devices -a add -d <node-number>:<bay-number>
The Add command connects the drive to the node. If the add command fails, use the following command to format drive
add failed.
- OneFS 8.0 以后 isi devices format <bay-number> --node-lnn <node_number>
- OneFS 8.0 以前 isi devices -a format -d <node-number>:<bay-number>
The Format command introduces the disk into the cluster.
some other situations
As shown in the figure below, the system will report a used drive from another cluster was inserted as a replacement. This is normal, because the replacement disk comes from a disassembled device.
Run the command isi status -n * * is the node number, you can list the information of this node
After performing add and format actions on these disks, the disks will enter the normal state.
What does it mean that Status is NONE?
After restarting the controller, the state becomes as follows:
The status of the disk changes to preparing, and then the disk keeps flashing.
It is estimated that it will take about 30 minutes. After the initialization is completed, it will become a normal healthy state, as shown below.
If the disk is force removed, that is, the hard disk is not removed through smartfail, then manually initiate
The action of smartfail, but the Lnum should be specified instead of the bay number