Link to home
Start Free TrialLog in
Avatar of Wean
WeanFlag for United Kingdom of Great Britain and Northern Ireland

asked on

How do I Move a Dell EMC FC 4500 San to a different server

I Have a Dell FC 4500 San and it looks as if we might have bought it second hand off of Noah, Its old and Massive and a waste of rack space, but I still have some respect for it and I would like to keep it running just for the IT Departments storage. What I do want to get rid off is the server it is attached to which is an old Dell 6650, and point it to our DFS Cluster instead.

I have connected the Dell FC4500 to two separate fiber switches and then setup the zoning to the HBA's on both nodes of the cluster, using SAN Surfer, I can see the Dell FC 4500, but I cannot see any LUN's and I have no idea how Im going to get them to be visible. My initial thoughts were to do it through San Surfer, but Ive tried and cant find where to configure it, then looking on the old Dell 6650, I found some software named navisphere, which i guess is the equivalent of Storage manger for IBM, and HP Array config utility for HP.

Is Navisphere what i need to present the LUNs to a certain WWN, or am I barking up the wrong tree?

Many thanks
Avatar of sathishchowdhry
sathishchowdhry
Flag of Ireland image

You would require to put in the Ip address of the one of the storage processor on a web browser (make sure you have java installed), login to the navisphere.

Right click on the array and goto connections, now you will see the new host is logged in as per the WWN of the HBA but not registered.
Register them first with the host name or the Ip address. Now you would require to add the host in the storage group to where LUn's are presented previously.

One the host in storage group, rescan the devices on the host side and you will be able to see the LUN's on the Host side
Avatar of Duncan Meyers
Navisphere is indeed what you need to set up the array - but with a twist. You should have a workstation set up as a Navisphere station somewhere on your network - the PE6650 may even be it. The host we're looking for will have the NaviGoverner service running on it. Once you locate that host, path to C:\Program Files\EMC\ManagementUI\<this will probably be 5 something>\WebContent and double-click on start.html. You'll be prompted for the IP address of the FC4500, then the username and password (admin/password as often as not). You can then follow satish's instructions, create a storage group and add the new host to the existing LUNs. Setting up a Navisphere workstation for an FC4500 is a bit of a palaver, so try and locate the system that's already configured and you'll save yourself some hassle.

The FC4500, as you note, was old when Noah was in short pants, but it had quite respectable performance and it's worth the trouble to fire up.
Avatar of Wean

ASKER

Cheers for your comments, I will give both of these ideas a go tomorrow and tell you how I get allong.
Avatar of Wean

ASKER

I have managed to find a Navisphere instaleed on the PE6650, (Navisphere Off-Array Management UI 6.26.0.2.23). And I have place an Ethernet Cable into SP1, and a console Cable into SP1 aswell. The Navisphere agent, and Nivisphere Govenor services are Running on the PE6650

When i clicked onto Navisphere Off-Array Management UI 6.26.0.2.23 it gives me 4 choices to loginto the SAN, SP IP, SP Hostname, Serial and NAT. I figured out via hyperterminal and the console  that the IP address of the SP-1 is 192.168.1.1, and after setting an IP I can ping the SP.

When i try Connect, It gives me an error of "Unable to connect to the provided IP Address and Port Number. Please verify IP Address and Port Number." Using 192.168.1.1 and both port options  "80/443" and "2162/2163"

If i try using the serial Connection, i Get a different error, which seems a little more hopefull "Unable to access security information" which makes me wonder if i need to set security somewhere first?

after having a chat to a colluge of mine he said that my predessor had tried to do the same thing Im trying to do and was unsucessfull, he also mentioned that the problem might have been the version of Navisphere that was installed onto the PE6650.

I can access the Console and im going to see if I can find any documentation on and commands i can type in to see user access or how the LUNS are setup.

If anyone has any ideas on how else i can try get into the SAN it would be much appreciated.

Many thanks

Try it on web browser with the IP address and see if you canget navisphere login. alternatevely check the security configuration using the http://<Ip Address of SP>/ setup
Avatar of Wean

ASKER

I placed the console cable into SP2 and found the ip of the SP by typing setlan, it is 192.168.1.5.

I cannot browse to either SP, im guessing this san might have needed an update for that functionality.

When i terminaled into the san using the console cable i found a few interesting things, by typing "?" i got a list of the commands i can use which i will attach as a file ("commands.txt") and i typed a few in, by using lustat, i could see a 597GB LUN, LUN 0 using FRU's 0 thu 9. And by typing Setdisk - L i could see each of the disks. But the best one of all was when i typed "getvlut".

As you can see below under "Currently Defined Initiators:" I can see the WWN's (Port and Node) for the 2 HBA's in the old PE6650.

HBA1: 200000E0.8B0606EB.210000E0.8B0606EB
HBA2: 200000E0.8B0621EB.210000E0.8B0621EB

Has anyone got any idea on how to change these?

Many thanks
fcli> getvlut
 
LUN REMAPPING: ENABLED
 
Array Name:
 
Default VA WWN: 60060160.00000000.00000000.00000002
 
Management Password CLEARED
 
Press any key to continue....
 
 
Currently Defined Initiators:
 
Idx Port VANum          World Wide Name             Assigned Name
--- ---- ----- ----------------------------------- -------------
  1    0    0  200000E0.8B0606EB.210000E0.8B0606EB @< 1>< 1>< F>\\.\SCSI2:0:2:0@
< 1>< 2>#lt-srv-fs01.tlt.domain.co.uk@< 1>< 3>< B>10.10.XX.XX@< 1>< 4>< 6
>ql2300@< 1>< 5><1C>QLogic Fibre Channel Adapter@< 1>< 6>< 6>QLogic@< 1>< 7>$f55
4aeee-4245-40aa-9870-ce9bdf06f618
  2    1    0  200000E0.8B0621EB.210000E0.8B0621EB @< 1>< 1>< F>\\.\SCSI1:0:2:0@
< 1>< 2>#lt-srv-fs01.tlt.domain.co.uk@< 1>< 3>< B>10.10.XX.XX@< 1>< 4>< 6
>ql2300@< 1>< 5><1C>QLogic Fibre Channel Adapter@< 1>< 6>< 6>QLogic@< 1>< 7>$f55
4aeee-4245-40aa-9870-ce9bdf06f618
 
Press any key to continue....
 
 
Currently Defined Virtual Arrays:
 
Num           World Wide Name                   Assigned Name
--- ----------------------------------- ---------------------------
  0 60060160.040A3196.A103BD90.24764104 SG_LT-SRV-SP01
 40 60060160.00000000.00000000.00000000 ~null
 41 60060160.00000000.00000000.00000002 ~management
 42 60060160.00000000.00000000.00000001 ~physical
 
 
Current VLU/FLU Mapppings:
 
VANum                 VLU/FLU Pairs (in hex)
----- ---------------------------------------------------
   0  0:0
  42  (all VLUs=FLUs)

Open in new window

commands.txt
Avatar of Wean

ASKER

I think i might have figured it out, but has anyone used the bind command using the console? or am i better off trying to get this done via navisphere?

Before i try it i want to make a full backup of the data, so im going to plug the fiber into the old PE6650 and do a full backup, then im going to plug the fiber back into the switch and give this BIND ALL command a go.

Im not sure what options to use either. If anyone feels im going to turn this SAN into a brick please tell me. Im not sure if i will be recreating a LUN here or just binding the current lun, either way it dosnt matter much after the backup, cause i can always do a restore to the new LUN if i destroy the current one.

Cheers
Usage: bind -h
       bind raid_type disk_names <options>
       bind all <options>
raid_type:r0 / r1 / r3 / r5 / r1_0 / id / hs / all
disk_names: e_d or e.d:'e' is enclosure address in hexidecimal(0-B)
            'd' is slot address(0-9) example:1.2 0_3
options:
   -a <1/0> enable / disable auto-trespass, default 0
   -c <none/read/write/rw> cache flags, default: none
       no-caching/read-caching/write-caching/read-and-write-caching
   -nm1    setup Raid-3 unit as N minus 1 and go
   -r <#>  rebuild time (in hours). default 4
   -s <#>  stripe size: number of sectors per element (disk)
           default: 1(RAID-3),128(non-RAID-3)
   -u <#>  logical unit number, default: -1(system determines the LUN)
   -z <#>  stripe-count. range: >=0  default: 0 (maximum number available)
   -addr <addr>  specify a logical address offset(hex) for the new LUN
   -ff     use first fit strategy for computing lun's address offset
   -g <#>  raid group, can be used instead of disk-names
   -L2d <#> amount of disk capacity for the L2 cache (in blocks)
   -L2m <#> amount of memory for the map in the L2 cache (in Megabytes)
 
Press any key to continue....
   -zm <#> LUN capacity, in Megabytes. range: >=1   (short-bind)
   -zg <#> LUN capacity, in Gigabytes. range: >=1   (short-bind)
   -repeat <#> bind multiple LUNs; do not use with '-u'
   -va <#> specify the virtual array number.  Default is physical
   -bm <#> bind mode: 1-FLU, 2-VLU 2.  Default is VLU
   -ab <#> align bind offset in sectors
   -ndb <password> nondestructive bind (do not zero LUN)

Open in new window

Unlike the later FC400 and CX arrays, the FC4500 does not have a web server built in - therefore Navisphere must be run from the PE6650. The IP address you've found (192.168.1.1) is the same on both SPs and is used for a PPP connection to do the initial configuration of the array.

The version of you are runniong on the 6650 is much, much, much later than the version of code the FC4500 is running. Your colleague may be correct - that the version 26 won't support the old kit - however, it's worth investigating a little further. Try logging into Navisphere locally using 127.0.0.1. You may be prompted to set up a domain (and here I'm scratching as it's been a lo-o-o-o-ong time since I've done this) and you'll see a dialogue box that has an option for IP addresses or subnets to scan - add the FC4500 SP IP address here. The Navisphere help files have quite a bit of information on this - or you could log a call with EMC support for some assistance

The bind all command won't do a whole lot as you don't appear to have any available space. You'd have to destroy the existing LUN first.

Finally, commands.txt contains your fully-qualified domain name. You may want to ask the site admins to remove the file.
Do you still conider to run production on this hardware as Iam aware it is really hard to find parts for replacement and will be hard to get the data back
Avatar of Wean

ASKER

Cheers meyersd,

After doing some research ive decided to reinstall Navisphere, the only version i can get my hands on is version 6.

sathishchowdhry, this wont be used for production, it will just be a file store for our department only. I have too much respect for it to see it being crushed, although it realy should be decomissioned.

This is my thought plan and I'm going to give it a go today. Im not sure whether to change the WWN on the curent PE6650, and then set this all up agin on the DFS Cluster, or to just go straight ahead and set it up on the Cluster, and switch the PE6650 off.


Im going to Install the "Management Server" software on the Both Cluster Nodes.
      (I think I will also have to install the "Managment UIs" on both nodes)
      It says to reboot after.  :-/

1b)      Configure the "Management Server" software on both nodes
      http://localhost/setup
      leave all settings as default and select Apply settings.

1c)      point a browser to the local IP of the Management Host on both Nodes , it should open up the Navisphere GUI.
      
        login as "admin / password"
      click File|Login and login as above.
      I should see a "Storage Systems" icon - and nothing else.

1d)      File|Setup Domain|SelectMaster
      Select the Management server's IP address (on each Node)

2)      Now we move the the "StorageHost".wich will also be the same two nodes)

2a)      Install the NaviAgent on each node
      Set the "agent.config" file to have the following
            clarDescr
            clarContact
            poll 30
            baud 9600
            event log 2048
            user administrator@(ip of Clustered virtual IP)
            user system@(ip of Clustered virtual IP)
      Savefile
2b)      Restart "NavisphereAGent" service.

3)        Back on the Managment Server
3a)      In the NaviGUI, Tools|PortalConfiguration
      Right click Portals, and select "Add Portal".
      Highlight the ManagementServer and move the right-hand column.
      The Managementserver should now be listed under Portals.

3b)      RightClick the ManagmentServer and select "Add Storage System"
      In "Select Systems" put the IP address of the "StorageHost" (step2)
      Hopefully the StorageHost will appear, and I can move it to the right-hand column.
      The StorageHost should be be listed under the ManageHost in the PortalConfig
      Click OK to close Portal Config.

4)      In the Hosts Tab, right click StorageHost, select Properties.
      In the Properties window click the Agent Tab.
      Make sure "AutoDetect" is ticked.

4a)      Restart the "Navisphere Agent" service on the StorageHost.

Hopfully my Mindset is correct here,

I will let you all know how it goes.

Oh and thanks meyersd for telling me about the command.txt file (Schoolboy Error)

>Im going to Install the "Management Server" software on the Both Cluster Nodes.
 >     (I think I will also have to install the "Managment UIs" on both nodes)
 >     It says to reboot after.  :-/
Rather than install it on the cluster nodes, install it on  workstation (even your own PC. As long as you have IP connectivity to the array, you can install it anywhere. It isn't a resource-hungry service. If you do install it on the cluster, you only need it on one node and it doesn't need resilience. The array keeps merrily working whether or not the Navisphere server is present. Yes - you'll also need to install the UIs.



>2a)      Install the NaviAgent on each node
>      Set the "agent.config" file to have the following
>            clarDescr
>            clarContact
>            poll 30
>           baud 9600
>          event log 2048
>            user administrator@(ip of Clustered virtual IP)
>            user system@(ip of Clustered virtual IP)
>      Savefile

You should have two entries:
 user system@(ip address of SPA)
 user system@(ip address of SPB)

Delete:   user administrator@(ip of Clustered virtual IP)

Otherwise you are spot-on.

Avatar of Wean

ASKER

Cheers meyersd,

I have changed the agent.config as follows,

clarDescr
clarContact
poll 60
baud 9600
eventlog 2048
user system@192.168.1.1
user system@192.168.1.5

I will sitback and see if it picks anything up.

Avatar of Wean

ASKER

AArrgh,

I think i have the wrong version of navisphere, I,m getting myself so wound up over 1/2 a Terrabyte! ]

The San is now back on the old PE6550, I can see the LUN on SANSurfer, and windows can see the drive, but i cant get naisphere to see the SAN at all. the agent is running on the PE6550 and the agent.config is as above. I can ping and telnet into both ipaddress's

Tell me its the wrong version of navisphere so i can take a baseball bat into the server room and settle a score!
ASKER CERTIFIED SOLUTION
Avatar of Duncan Meyers
Duncan Meyers
Flag of Australia image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of Wean

ASKER

Well, looks like its the baseball bat!

Thanks for all your help meyersd. if i do find the voodo doll controling the SAN somewhere in the server room and manage to reserect this thing i'll be sure to tell you how.

Cheers