3PAR Procedures - Replace Disk, Remove Disk & Change Disk Position - Rebalance

In the first Post of the year I will show you three procedures regarding disk failure, remove a disk from the CPG and rebalance disk position in a 3PAR array.

• Replace a failed disk in 3PAR
• Objective: Remove failed disk and replace with a new one
• Remove disks from the CPG and rebalance the data
• Objective: Remove a disk from the 3PAR. The disk will be in New state and you can use it in another 3PAR
• Change disk position in 3PAR so you can rebalance data in the array
• Objective: Maximize performance of your array when you upgrade the 3PAR with new cages or disks

Let's see how to do this three procedures.

Replace a failed disk

Note: Don't forget to put the 3PAR in maintenance mode (in the Service Processor)

1. Show the failed or degraded physical disks
    showpd -failed -degraded

2. Take note of the failed/degraded disk ID

3. Start servicemag command to inform the system to evacuate all the chunklets from a drive so that it is ready for service.
   servicemag start -pdid ID

4. Check the servicemag status. It can take some time to remove all the chunklets
   servicemag status

5. Check if there is no data on the drive. All the columns other than size and failed are zero
   showpd -space ID

6. Replace the failed/degraded physical disk

7. Check the servicemag status for the rebuilding. It can take several hours
   servicemag status

8. Once this is complete your disk must have state normal and green
   showpd -state 

If you have more that one failed disk, repeat the procedures above.

Remove disk from CPG

Note: Don't forget to put the 3PAR in maintenance mode (in the Service Processor)

1. Verify the disk ID and take note
    showpd -s

2. Stop data on the drive
    setpd ldallof off ID

3. Make sure the disk is now in Degraded state
    showpd -s

4. Vacate all data from PD (-f) no confirmation (-vacate) move of data is permanent
    movepdspare -f -vacate -nowait ID 

5. Check if there is remaining chunklets to be moved
    showpch -mov ID

6. If with the previous command you get error, move chunklets manually (X=ID, Y=chunklet). Repeat till disk is empty
    movech -perm -ovrd X:Y

7. Check if there is remaining chunklets to be moved
    showpdch -mov ID

8. If there is no more data chunklets to be moved, let's move on to spare chunklets
    showpdch -spr ID

9. Remove all the Spare chunklets from the PD
    removespare ID:a

10. Dismiss the disk from the 3PAR, Once the Disk shows up as "new" status, it can be removed
    dismisspd ID

11. Remove the vacated disk from the 3PAR

12. Repeat the procedures above for all the disks that you want to remove

13.Rebalance the 3PAR with tunesys
    tunesys -chunkpct 1 
    tunesys -cpg

14.Check if 3PAR is OK
   admithw -checkonly   
   checkhealth -detail

Change disk position in 3PAR Cages

Note: Don't forget to put the 3PAR in maintenance mode (in the Service Processor)

1. Locate the source and destination drive cages
   locatecage

1. Relocate disk from the source to destination 
   relocatepd start -f <src_cage>:<src_slot> <dst_cage>:<dst_slot>

2. To proceed with the "relocatepd" operation:

• Remove the drive from the indicated source location.
• Insert it into the indicated destination location.
• Resume the operation 
• relocatepd resume

3. To cancel the "relocatepd" operation:

• Leave the drive in place.
• Cancel the operation
• relocatepd cancel

In either case, the alert will be resolved automatically by the system, when the "relocatepd" command completes.
Let's see an example, where you want to rebalance all the disks to get more performance. If you have 3 cages with the following disks:

• 10*480GB SSD
• 24*450GB SAS
• 16*1.2TB SAS
• 8*600GB SAS

The best configuration according to Ninja3PAR is this:

 

If you have the disks like the configuration in the picture below, you have to do this procedure for all of them:

 

That's it. You have any questions or want to add some more stuff, please let me know.