VPLEX Provisioning

EZ-Provisioning

1) Claim storage and create virtual volumes

Option 1 – Use the Create Virtual Volumes wizard to automatically create a virtual volume and the underlying extent and device.  This option allows you to use unclaimed or claimed storage.  The wizard eliminates the separate steps required to claim storage volumes, create extents, and then create devices.  Launch the wizard from the Storage Volumes or the Storage Arrays screen.

  1. Provisioin Storage > cluster > Storage Volumes or Storage Arrays
  2. Create Virtual Volumes
Example: Metro config might have a Storage Volume in Cluster-1 SiteA_SAN1_LUN1000 which has a corresponding SV in Cluster-2 at SiteB_SAN1_LUN1000

Option 2 – Manually create the virtual volume on an existing top-level device.  Show Properties Links can be used to view the details for a device.  Alternatively, select one or more devices from the Devices screen and click Create Virtual Volumes.

  1. Provision Storage > cluster > Virtual Volumes.
  2. Click Create.
  3. Select device and click Add

2) Register Initiators

VPLEX automatically discovers initiators connected directly or via the FC fabric and populates the Initiators screen accordingly.  Once discovered, the initiators must be registered with VPLEX before they can be added to a storage view.  The registration process gives the PWWN a meaningful identifier such as the hosts DNS name.  If VPLEX fails to automatically discover an initiator it can be registered manually.  Note that each port of a host’s HBA/HCA must be registered as a separate initiator.

Initiator Host types can be selected in the Host Type column and should be selected before registration is complete since it cannot be changed afterwards except via the CLI.  Host types supported include: SUN-VCS, AIX, HPUX, RecoverPoint, or Default for use with any type not listed.

  1. Go to Provision Storage and select cluster > Initiators.
  2. Select the unregistered initiator and click Register.
  3. Type a meaningful name for the initiator or accept the one provided.
  4. Select a host type and click OK
  5. Close the dialogue after successful completion.

To manually register:

  1. Follow same steps as above except select Register Initiator Manually on the Initiators screen.
  2. Type a meaningful name for the Initiator and select the host type.
  3. Enter the PWWN and click OK

Note: if the PWWN does not display after registration an invalid PWWN may have been entered or there are zoning/cabling issues.  If there are no target ports listed for the initiator then it is not logged into the system, if they are listed then an invalid PWWN has been entered most likely.  If an invalid PWWN was entered just unregister the initiator and register again with correct info.

  1. To unregister, select the initiator in the Initiators screen and click OK.

3) Create Storage View

A storage view must be created in order to export storage.  Then VPLEX FE ports and virtual volumes need to be added to the view.  Virtual volumes are not visible until they are associated with ports and initiators.  The Create Storage View wizard creates a storage view adding all of the requisite items to the view and makes it active.  Once active, the storage view can start serving I/O to the virtual volumes.

Guidelines:

Create a separate storage view for each host accessing the same storage and then add the virtual volumes to each view.

At least two initiators per host are required for redundancy.

LUN numbers can be assigned automatically or manually by providing a range of LUN numbers to use.

  1. Go to Provision Storage and select cluster > Storage Views
  2. Click Create and follow the wizard
Example: A Metro config could have a vplex1_cluster in both sites each containing their corresponding Initiators, Ports and Virtual Volumes.

Advanced Provisioning

1) View available storage

  1. Go to Provision Storage select cluster > Storage Arrays

From here you can:

  • See name of discovered storage array
  • View the number of LUNs exported from each array (Logical Unit Count)
  • Determine if Auto Failover is enabled for active/passive or ALUA arrays
  • Check the Connectivity Status of the array to determine if at least one path to both of the controllers is available.
  • View array properties
  • Show Logical Units for the array or Rediscover the Array if the number of logical units is incorrect
  • Find an array by typing part of the name in the Find box
  • Show/Hide the properties links at the bottom of screen
  • Start the Create Virtual volumes wizard
  • Start the Claim storage wizard
  • Set the rebuild type of the array to thin

  a)   Rediscover the array to view recently zoned and masked storage that has been presented to VPLEX

    • Select the array from the drop down
    • Click the rediscover array button

2) Claim storage volumes

Storage volumes must be claimed before they can be used in a cluster.  Once claimed they can be used to create extents, devices and virtual volumes.  Before claiming storage volumes the storage arrays should be viewed to ensure that they’ve been discovered correctly with the appropriate number of unclaimed storage volumes.

Some arrays require a Name Mapping file in order to claim storage; VPLEX will automatically detect whether or not this is required.  If a storage array does require a Mapping File the wizard will prompt for a file upload.  Even if a Mapping File is not required it can still be used if customized naming is desired.  The default naming convention utilizes the SCSI VPD (Virtual Product Data).  Storage volumes may also be renamed after claiming them.

Example command to upload mapping file for CLARiiON/VNX:

NaviSECCLI -h <CLARiiON_IP> -user <user> –password <password> –scope 0 getlun -uid -name > <map_file_name>.txt

Note: File name being piped to must start with letter or the command will fail. Piped file will go to directory command run from. Vplex uses filename for its naming; using the same file name will ensure overwrite avoiding multiple files.

where, for example, map file contains:

LOGICAL UNIT NUMBER 1000

UID:                        60:06:01:60:1C:D0:2D:00:A4:A7:18:BB:76:FF:EE:22

Name                        LUN 1000

LOGICAL UNIT NUMBER 1001

UID:                        60:06:01:60:1C:D0:2D:00:A4:A7:18:BB:76:FF:EE:23

Name                        LUN 1001

 

…….. and so on for each LUN

Note that the filename used for the mapping file is prepended to the mapped storage but this name can be changed to a unique unused name after the mapping file has been imported.  For example, if mapping file is named dc1_vnx1.txt then each mapping will be dc1_vnx1_<volumename> by default.

Mapping file can be transferred via the GUI or by using the CLI workaround in section b) below.  CLI is mandatory in some cases with 3rd party arrays or when the GUI does not recognize the file.  Hint: Save the file in notepad as UTF8.

a)  To claim a single LUN:

  1. Open VPlexcli and cd to clusters/<clustername>/storage-elements/storage-arrays/<vendorarray>   #hint, use tab completion to list available clusters and arrays
  2. Run    ll logical-units 
  3. locate the VPDXXXX:XXXX... id that corresponds to the LUN you want to claim
  4. Choose an appropriate identifier and run the following command in the same directory as above to claim the volume.
  5. storage-volume claim -n <identifier> -d VPDXXX:XXXXXXXXXXXXXXXXXX...
  6. The volume is now available to create an extent.
b)  Workaround for 3rd party mapping or when GUI will not cooperate:
  1. Connect to VPLEX using WinSCP or another SCP capable client as admin or service from the system that has the mapping file stored locally.
  2. Change directory into /tmp at the root of the VPLEX file system in WinSCP.
  3. Copy the mapping file over to the /tmp directory from the local filesystem.
  4. Then ssh into the VPLEX server with admin or service account and enter the Vpexcli mode.
  5. Verify that the file has uploaded properly with cat /tmp/<mappingfilename>
  6. Change directory to clusters/<clustername>/storage-elements/storage-arrays/<vendorarray>
  7. Run storage-volume claimingwizard -f /tmp/<mappingfilename>
  8. The storage volumes are now available to create extents.

3) Create extents from storage volumes

The Create Extents wizard can be used to create extents for storage volumes; VPLEX will automatically create one extent the same size as the selected storage volume but a smaller size can be specified to reserve capacity for later use.  A total of 128 extents per storage volume is supported.

If a storage volume is larger than the desired virtual volume extents should be created with the desired storage volume size; avoid creating smaller extents and different RAID types to concatenate or stripe extents.

On the other hand, if the storage volume is smaller than the desired virtual volume, create a single extent per storage volume and then use devices to concatenate or stripe the extents into a larger device.  If a storage volume fails to create an extent investigate the status in the Storage volumes screen and troubleshoot as necessary.

  1. Go to Provision Storage > cluster > Extents.
  2. Click Create.
  3. Follow the wizard instructions.
  4. Alternatively, the Storage Volumes screen has a Create Extents link after selecting storage volumes to have wizard pre-populated with volumes to use.
Example: Metro config might be setup with Cluster-1 having extent_SiteA_SAN1_LUN1000_1 with corresponding extent at second site in Cluster-2 named extent_SiteB_SAN1_LUN1000_1

4) Create devices from extents

The RAID-0 (higher depth generally gives higher performance), RAID-1, and RAID-C  (combines the extents in a linear fashion – concatenation) type devices are created with the Create Devices wizard.  Additionally, 1:1 Mapping of extents to devices can be performed.  Each RAID type supports a maximum of 64 extents.  The CLI provides more advanced options for creating devices; for example, CLI may be used to create a device with different device types.

  1. Go to Provision Storage > cluster > Devices.
  2. Click Create.
  3. Follow the wizard.

If you have any questions on above steps, you may reach out to me at Facebook Subash Krishnan or at Linkedin Subash krishnan 

Leave a comment