How to physically move bound drives from one DAE enclosure to another in
VNX2 with MCx
Article Number:000190186 Version:3
Key Information
Audience: Level 30 = Customers Original Create Date: Tue Jul 22 16:40:53 GMT 2014 Article Type: How To
Channels: Customer , Internal App First Published: Tue Jul 22 18:03:43 GMT 2014 Validation Status: Technically Approved
Originally Created By: Daniel Nowlan Last Modified: Tue Jul 22 19:20:50 GMT 2014 Last Published: Tue Jul 22 19:20:50 GMT 2014
Summary: With the drive mobility feature of MCx, it is possible to move bound drives to another location of the array if desired. This KB article will assist with
understanding the details around this feature and the actual methodology used..
Article Content
Goal: This procedure was created to describe the method to physically move bound drives from one enclosure to another enclosure in VNX2 with MCx (Release 33.051 and higher.)
Issue: Usage:
With this procedure a complete RAID Group (RG) can be moved from one enclosure to another enclosure to help with performance or for moving
drives for maintenance reasons. In reality, in VNX2, drives associated with a RG can be scattered around the back end on different buses and in
different enclosures even after binding.
IMPORTANT information regarding moving a failed drive that was rebuilt to another drive and then moving that rebuilt drive back to its original physical
location:
In VNX2, when a drive fails in a RG, after 5 minutes, a new unbound drive will be selected and the RG will rebuild to that selected drive. This
procedure can be used to move that rebuilt drive to the position of the failed drive lieu of doing any data copy. There is no concept of equalize in
VNX2, once a failed drive is rebuilt to another drive, that new drive is permanently a member of the RAID Group. If the failed drive is replaced, there is
no equalization done.
Environment: Product: VNX5200
Product: VNX5400
Product: VNX5600
Product: VNX5800
Product: VNX7600
Product: VNX8000
SW: VNX-Block-Operating-Environment-05.33.000.5.051 or higher
Cause: N/A
Change: N/A
Resolution: Restrictions:
This procedure cannot be used to physically move any of the four system drives (0_0_0, 0_0_1, 0_0_2 and 0_0_3). These drives cannot be moved and must remain in Enclosure 0_0 in the
exact order that they are found in the array.
Preparation:
First get spcollects and Triiage them. Using the LUSTAT section of the Analysis file determine and record the RAID Group(s) or drives that you plan to physically remove and replace in another
enclosure and/or slot.
If the RAID Group that you are moving a drive from is broken or degraded then you must first repair that RAID Group prior to executing this procedure.
Procedure:
Once you have determined the drives and/or RAID Group you physically want to relocate from one enclosure (or slot) to another do the following steps:
1. Select the first drive that needs to be removed and insure the RAID Group is not broken. You can check the status of the RAID Group using the following Naviseccli command:
Naviseccli getrg –messner <RG #>
The Unisphere GUI can be used to monitor the drive activity (Select Storage>StoragePools, select Pools or RAID Groups. Select the Pool or RAID Group in the selection pane and select
disks. This will show you the BED position of the disks in the Pool or RAID Group).
2. If the RG is not broken or degraded then the first drive can be moved. Remember in Release 33 you have 5 minutes to re-insert a drive when removed prior to a hot spare being invoked
(write logging will be done for the time the drive is not present in the array up to 5 minutes). Only a single drive per RAID Group can be moved at a time.
3. Place the drive back in the array in the enclosure and slot you choose to move the drive and/or RG to. Remember, you have 5 minutes to re-insert the drive once removed to avoid a HS
being invoked.
4. Once the drive has been inserted, check status of the drive using the fbecli command mentioned in step 1. The drive will may initially report rebuilding but once the rebuild is complete the
drive will report its BED position as the new position. Once this is complete this drive move is complete.
5. If more drives in the RAID Group need to be moved follow the steps again. Always remember, only a single drive from each RAID Group can be moved at a time.
Example of FBECLI outputs. Started with drive in slot 2_0_21 and moved to slot 0_0_23.
Initial configuration of RG 998
RG 0x165 :357 998 RAID5 READY 352 353 354 355 356 2_0_21 2_0_20 2_0_19 2_0_18 2_0_17
Configuration after moving drive 2_0_21 to BED 0_0_23
RG 0x165 :357 998 RAID5 READY 352 353 354 355 356 0_0_23 (REB:0%) 2_0_20 2_0_19 2_0_18 2_0_17
Configuration after rebuild of write log done
RG 0x165 :357 998 RAID5 READY 352 353 354 355 356 0_0_23 2_0_20 2_0_19 2_0_18 2_0_17
Article Metadata
Product: VNX2 Series, VNX5200, VNX5400, VNX5600, VNX5800, VNX7600, VNX8000
Shared: Yes
Originally Created By: Daniel Nowlan
