Mapping all drives Individually to Windows 10

  • 119 Views
  • Last Post 09 October 2018
  • Topic Is Solved
Mike Duncan posted this 26 August 2018

Hello all,  I've been tasked with mapping all of the drives in an Ex30 (with five Jx30's 96 drives total) to a Windows 10 machine via 8G fibre.  I'll be connecting the two fibre cables from the Ex30 (Point-to-Point) to the Windows machines ATTO 82EN fibre card.  All data on the storage is archived to LTO. 

From there I'm not sure what happens next to get the drives mapped and showing up individually in Windows 10 Disk Managment. 

If anyone is willing to share an outline or step by step guide on how to make this happen that would be great.

Thanks

mike_d

Order By: Standard | Latest | Votes
arunkumar b posted this 27 August 2018

Hi Mike,

Please follow the below steps for LUN Mapping.

Adding an FC Initiator:

You must add an initiator to the VTrak’s initiator list in order to use the initiator to

create a LUN for your logical drive.

To add a Fibre Channel initiator from a list:

1. Click the Device tab.

2. Click the FC Management icon.

3. Click the Initiators on Fabric tab.

4. Check the box to the left of the initiator you want to add.

5. Click the Add to Initiator List button.

The initiator is added, and its check box grays out.

 

Adding a LUN Map:

LUN masking must be enabled in order to assign LUNs to logical drives.

Disabling LUN masking allows all initiators to access all LUNs in your data

storage.

 

To add a LUN map:

1. Click the AdminTool tab.

2. Click the LUN Mapping & Masking icon.

3. Click the LUN Mapping button.

   The first LUN Mapping screen appears.

4. Choose an initiator from the Select Initiator dropdown list.

5. Click the Next button.

   The second LUN Mapping screen appears.

6. Drag a logical drive from the Logical Drive list and drop it onto the Initiator

   list.

7. Click the Next button.

   The final LUN Mapping screen appears showing the initiator and LUN map.

8. Click the Submit button.

   The new LUN map is created.

Now go to Windows 10 disk management and scan the disk to see if the volume is visible

 

Regards

Arun

Promise Tech Support

Richard Oettinger posted this 27 August 2018

Hi Mike,

Do you wish to have 96 individual disks on one Windows system?

To do that you need to create one Disk Array using one drive, then create a Logical Drive with it - 96 times.

Grab the product manual and check out the section starting at page 151:

https://www.promise.com/DownloadFile.aspx?DownloadFileUID=1151

Because there will only be one disk in each Disk Array, you can only create a RAID0 Logical Drive, which means that you will have zero protection against a drive failure...

 

R P posted this 27 August 2018

Hi Mike,

Follow Richard's advice making 1 RAID0 array/LUN per drive, but there are some cautions.

You cannot mount all the drives to a drive letter, Windows drive letters go from 'a' to 'z' and then they fall off the end of the alphabet. Also, 'a' and 'b' are reserved for floppies and 'c' is probably your boot drive so you can only mount 23 drives to a drive letter, unless you have a DVD or any other devices mounted to a drive letter, in which case less than 23.

But, modern windows also allows you to mount the drive to a folder (a lot like Unix), so you can create 96 folders with whatever naming scheme seems appropriate and mount the LUNs to the folders. This 'superuser' article shows how.

https://superuser.com/questions/536886/what-happens-when-windows-runs-out-of-drive-letters

If you are on a SAN, then you should LUNmap, but if you are direct connecting the Ex30 to your windows server, as seems to be the case, then LUNmapping won't be necessary,

Mike Duncan posted this 03 September 2018

Hello all,

I see now that I need to connect as DAS, add the Initiator and and create a LUN with 96 drive in it.  When the new LUN map is started.  How much time should I estimate for the initialization?  96x2TB drives, 48 hour?

Thanks to all for the insight.  I'll be making the changes in a couple of weeks and I'll keep you posted on how it turns out.

mike_d

Richard Oettinger posted this 04 September 2018

If the unit is connected to one server you do not need to add the initiator. With LUN Mapping disabled, any Logical Drives you create will be visible to the server.

We - and other vendors - dont typcially recommend creating arrays with 96 drives. The rebuild time for one failed drive is greatly extended as every drive has to be read to create the parity for the replacement.

A rule of thumb would be to create a RAID6 LD using all the drives in an enclosure, then you could either have five drives available to Windows or create a spanned volume...

 

Mike Duncan posted this 02 October 2018

The storage was connected to multiple MacPro's working as the MDC's.  The Mac's are now gone and the storage plan has changed.  I've got the Windows 10 sytem racked and the VTrak conneted as DAS storage (P2P).  The storage was previously on the fibre switch.  I've yet to take a look at the webpam from the new computer.  The current LUNs are not visible, but I didn't expect them to be.  

I was not onboard with one huge LUN either and have broken it down to three LUNs.  Two 40TB LUNs and one with the remainding space.  This best matches how it was previously configured.  I'm hoping this can be configured via the webpam. 

The level of the Wizard will allow me to define the LUN sizes, RIAD6 with hot spares?

 

 

PROMISE Technology Inc. posted this 02 October 2018

Hi Mike,

The current LUNs are not visible, but I didn't expect them to be.

If lunmapping was previously enabled, then the LUNs won't be visible as they are mapped to the MAC Pros. Unless I am reading you wrong, the storage is direct attached now, and if that is the case you should probably just disable lunmapping. This is a setting in the controller page of WebPAM.

I was not onboard with one huge LUN either and have broken it down to three LUNs.  Two 40TB LUNs and one with the remainding space.  This best matches how it was previously configured.

The previous configuration was for a Stornext/Xsan filesystem with a small LUN for metadata and 2 data luns for better performance. There is no benefit to using this configuration with a DAS. But it will work just fine.

I'm hoping this can be configured via the webpam. 

Yes, and the CLI can be used as well.

The level of the Wizard will allow me to define the LUN sizes, RIAD6 with hot spares?

You will have to use the Advanced level of the wizard to create your configuration, although it's easy enough to create arrays and LUNs from the storage page.

Mike Duncan posted this 09 October 2018

Hello Richard and anyone willing for join this adventure in VTrak Win 10 configuration.    

I'm going to rewind a bit.  I was not prepared for the install and decided to do a per incident call with Promise Support to get things working quickly and with some confidence.  The per incident support has been a complete failure.  After working with them for a week while onsite they are now telling me VTrak is not supported on Windows 10.  If this were the case why didn't they tell this when I outlined the gear on the first phone call.  I'm not happy with Promise support and I'll leave it at that.

I need a new plan of attack.  I had a week onsite to get things working, but the old gear wasn't up to par.  Controller 1 works enough to make a WebPam connection and the FC ports are all offline.  Controller 2 is dead.  I did flash Controller 1 to get it current, but this didn't help it.

If I can obtain two new controllers for this Ex30 is it possible to map (I'm using the word map becuase I don't know what to call it on Windows) LUNs from the Vtrak to a Windows 10 computer?

If possible I then need an outline as to how to configure it.  

What I think happens:

Setup the gear in DAS as shown on page 38 of the VTrak_SBB_PM_v1.1_final.pdf

Use WebPam to create the LUNs. Three LUNs all Raid5 (40TB, 40TB & remaining space) I have a huge problem at this point because nothing but load a script is avaialble in the wizard.  All other Wizard options are greyed out.  I'm hoping this is due the bad controller. 

If I can even make new LUNs the next step would be to install Perfect Path or try the CLU Quick Start.  I know Perfect Path will not install on Windows 10 (I tried multple times and it never installed).  I ran out of time onsite before I could try CLU Quick Start.  

And that this is where I'm stuck.  I'm sure there are more steps and hopefully you don't mind filling in the blanks for me.

Thanks

mike_d 

 

 

P B posted this 09 October 2018

Mike,

I see that you have an active case in our CRM. I looked at the logs and you may want to understand your task of execution a bit better.

1. You're requesting on how to migrate a SAN that was being used on Xsan to a Windows environment.
2. In your logs, you have a total of (17) arrays and (17) LUNs:

[Disk Array Info]

===============================================================================
DaId Alias     OpStatus      CfgCapacity FreeCapacity   MaxContiguousCapacity 
===============================================================================
0    MDC       PDM Running   4 TB        0 Byte         0 Byte          
1    Data1     OK            12 TB       0 Byte         0 Byte          
2    Data2     OK            12 TB       0 Byte         0 Byte          
3    Data3     OK            12 TB       0 Byte         0 Byte          
4    Data4     OK            12 TB       0 Byte         0 Byte          
5    Scratch1  OK            6 TB        0 Byte         0 Byte          
6    Data 5    OK            12 TB       0 Byte         0 Byte          
7    Data 6    OK            12 TB       0 Byte         0 Byte          
8    Scratch 2 OK            6 TB        0 Byte         0 Byte          
9    DATA7     PDM Running   12 TB       0 Byte         0 Byte          
10             OK            12 TB       0 Byte         0 Byte          
11             OK            6 TB        0 Byte         0 Byte          
12   Data9     OK            14 TB       0 Byte         0 Byte          
13   Data10    OK            14 TB       0 Byte         0 Byte          
14   Data11    OK            14 TB       0 Byte         0 Byte          
15   Data12    OK            14 TB       0 Byte         0 Byte          
16   META_SAN3 OK            4 TB        0 Byte         0 Byte          

[Logdrv Info]

=============================================================================== LdId Alias OpStatus Capacity Stripe RAID CachePolicy SYNCed =============================================================================== 0 MDC OK 1.99 TB 64 KB RAID1 RCache/WBack Yes 1 Data1 OK 9.99 TB 1 MB RAID5 RAhead/WBack Yes 2 Data2 OK 9.99 TB 1 MB RAID5 RAhead/WBack Yes 3 Data3 OK 9.99 TB 1 MB RAID5 RAhead/WBack Yes 4 Data4 OK 9.99 TB 1 MB RAID5 RAhead/WBack Yes 5 Scratch1 OK 3.99 TB 64 KB RAID5 RAhead/WBack Yes 6 Data 5 OK 9.99 TB 64 KB RAID5 RAhead/WBack Yes 7 Data 6 OK 9.99 TB 64 KB RAID5 RAhead/WBack Yes 8 Scratch 2 OK 3.99 TB 64 KB RAID5 RAhead/WBack Yes 9 Data 7 OK 9.99 TB 64 KB RAID5 RAhead/WBack Yes 10 Data 8 OK 9.99 TB 64 KB RAID5 RAhead/WBack Yes 11 Scratch 3 OK 3.99 TB 64 KB RAID5 RAhead/WBack Yes 12 Data9 OK 11.99 TB 64 KB RAID5 RAhead/WBack Yes 13 Data10 OK 11.99 TB 64 KB RAID5 RAhead/WBack Yes 14 Data11 OK 11.99 TB 64 KB RAID5 RAhead/WBack Yes 15 Data12 OK 11.99 TB 64 KB RAID5 RAhead/WBack Yes 16 META_SAN3 OK 1.99 TB 64 KB RAID1 RAhead/WBack Yes

3. If you want to delete the (17) arrays & (17) LUNs and start fresh (something that is more adequate for a Windows environment), you'll need to delete the arrays and the LUNs. (Deleting the arrays will delete the LUNs that pertain to each array).

#To delete via CLI#
#This will delete all 17 arrays#

array -a del -d0~16

3A. You can also delete all these (17) arrays via the WebPAM GUI as well.

  • Go to Disk Array
  • Select the Disk Array you want to delete and select Delete (You'll be prompted to type CONFIRM prior to deleting):


I have a huge problem at this point because nothing but load a script is avaialble in the wizard.  All other Wizard options are greyed out.

  • Once you've deleted you arrays (and the LUNs that pertain to these arrays), you can use the wizard to re-configure the RAID head (E630) and all the JBODs as desired. You stated that the wizard is grey, the wizard is grey because the entire subsystem is still configured and there are no more drives to configure. The drives have to be unconfigured in order to re-configure them. All 96 drives are still configured.

4. CTRL1 only has (1) 4Gb SFP installed and it's offline:

[FC SFP Info]

-------------------------------------------------------------------------------
CtrlId: 1                              Port: 1
Transceiver: SFP                       Connector: LC
TransCode: 00-00-00-00-20-40-0c-15     SerialEncoding: 8B10B
BitRate: 4 Gb/s                        ManufacturingDateCode: 110410  
LinkLength_9/125um: 0 m                LinkLength_50/125um: 150 m
LinkLength_62.5/125um: 70 m            VendorName: FINISAR CORP.   
VendorOUI: 00-90-65                    VendorPartNum: FTLF8524P2BNL   
VendorRevision: A                      VendorSerialNum: PKF558Y         

-------------------------------------------------------------------------------
Controller 1 Port 2 : No SFP found

-------------------------------------------------------------------------------
Controller 1 Port 3 : No SFP found

-------------------------------------------------------------------------------
Controller 1 Port 4 : No SFP found

-------------------------------------------------------------------------------

[FC Port Info] ------------------------------------------------------------------------------- CtrlId: 1 Port: 1 State: Offline PortIdentifier: 0xffffff TopologyAttached: N/A WWPN: 26-03-00-01-55-51-ca-bb Alias WWPN: N/A FabricWWNN: N/A FabricWWPN: N/A CurrentSpeed: Unknown LinkType: Unknown ConfiguredLinkSpeed: Auto ConfiguredTopology: Auto HardALPA: 224 ------------------------------------------------------------------------------- CtrlId: 1 Port: 2 State: Offline PortIdentifier: 0xffffff TopologyAttached: N/A WWPN: 26-02-00-01-55-51-ca-bb Alias WWPN: N/A FabricWWNN: N/A FabricWWPN: N/A CurrentSpeed: Unknown LinkType: Unknown ConfiguredLinkSpeed: Auto ConfiguredTopology: Auto HardALPA: 225 ------------------------------------------------------------------------------- CtrlId: 1 Port: 3 State: Offline PortIdentifier: 0xffffff TopologyAttached: N/A WWPN: 26-01-00-01-55-51-ca-bb Alias WWPN: N/A FabricWWNN: N/A FabricWWPN: N/A CurrentSpeed: Unknown LinkType: Unknown ConfiguredLinkSpeed: Auto ConfiguredTopology: Auto HardALPA: 226 ------------------------------------------------------------------------------- CtrlId: 1 Port: 4 State: Offline PortIdentifier: 0xffffff TopologyAttached: N/A WWPN: 26-00-00-01-55-51-ca-bb Alias WWPN: N/A FabricWWNN: N/A FabricWWPN: N/A CurrentSpeed: Unknown LinkType: Unknown ConfiguredLinkSpeed: Auto ConfiguredTopology: Auto HardALPA: 228

4A. Either go through you FC switch or you can go to a P2P (Point-to-Point)

  • P2P = Windows HBA > (1) initiator port to each controller's FC target port. (If you had (2) controllers active)
  • Since you have (1) controller that is online, and we have not determined the root cause of controller 2, you can still use controller [1] to create the data volume(s) as required. (Make sure you configure your arrays and LUNs appropriately; refer to Step 3 & 3A)
  • Make sure you're using appropriate FC cables and SFPs:

If I can even make new LUNs the next step would be to install Perfect Path or try the CLU Quick Start.  I know Perfect Path will not install on Windows 10 (I tried multple times and it never installed).

5. The Promise PerfectPath was specifically designed for Windows Server operating systems; not standard versions of Windows (7,8, 8.1 or 10).

5A. Customers that deploy VTraks Ex30 or Ex50s in Windows environments to provide file-level services (i.e. SMB), will typically use a Windows Server version and not a consumer version of Windows. That being said, using Windows 10 with the Ex30 is completely possible; as long as the prerequisites are taken care of.

5B. Below are the qualified Windows operating systems that you can use our PerfectPath driver v4.1.0.12:

  • PerfectPath Release Notes V4.1.0.12
  • If you can use an enterprise version of Windows (2008, 2012), that would be more desirable. (Don't use Windows Server 2016 as we currently don't support it.)

6. The only way to know what's wrong with controller 2 is by plugging in a serial and capturing what the serial console outputs.

7. Promise also offers Professional Services contracts in the event you would like us to come on site.

Mike Duncan posted this 09 October 2018

Major hitch in the giddy up.  Ex30 EOL'ed back in May 2018 and there are no more replacement parts. 

Both of the controllers I've been working with are toast.  No replacement parts nails this Ex30's coffin shut. 

The best advice I can give to anyone who has a VTrak Ex30 that is still running is Archive & Backup daily.

I lost zero data because we fully archived & backup daily to LTO.

Thank you to everyone who chimed in.

P B it would have made the Per Incident call worthwhile had I been able to talk with you last week. 

 

mike_d

Close