Tag Archives: VNX

VNX DPE / DAE SAS cabling

A quick note to clarify SAS cabling between EMC VNX and extra enclosures or DAE’s as they are called.

When setting up a VNX with only the base DPE enclosure, there’s no need for cabling.

When adding the first and the second enclosure, you need a special kind of SAS cable, called the “mini-SAS to mini-SAS HD cables”.

When adding a third enclosure or more, you need a “mini-SAS to mini-SAS cable”.



The following is an overview on the partnumbers for both situations:VNXSAS



Quicksetup EMC Connectrix DS-300B



The EMC® Connectrix® DS-300B provides a cost-effective, flexible foundation for entry-level SANs and can serve as an edge switch for core-to-edge SAN environments.

24 ports in 8-port increments through Ports-on-Demand licenses at 8, 16, and 24 universal (E, F, M, FL, or N) ports

Auto-sensing of 2, 4, and 8 Gb/s port speeds

Basic setup:

  1. connect your laptop via the serial-utp connection to the serial-utp port on the switch (IOIOI)
  2. open a putty/serial connection
  3. login with default credentials: admin/password
  4. setup network with commands (ipaddrset, ipaddrshow):
  5. setup date/time:
    tstimezone –interactive
    date “mmddHHMMyy”
  6. switch names:
    switchname “name”
    chassisname “name”
  7. licensing:
    licenseadd “key”
  8. ethernet cabling

Now you should be able to use your browser/java to connect to the GUI.


In the GUI, you need to open the “Zone Administration” page.
In most of cases, I’m doing fabric-based soft zoning. I’m almost never doing port zoning because things will go wrong when someone misplaces a fibre cable.
Recommended approach is using “Single-HBA” zoning, so Each zone created has only one server-HBA (initiator) in the zone together with all of the target storage devices.

For instance

1. create Aliases

A zone alias is a familiar name assigned to a device or a group of devices.example: alias “SPA_P0” = 50:06:01:60:bb:60:00:56 (EMC VNX Storage Processor A, Port 0)
60=SP A Port 0
61=SP A Port 1
68=SP B Port 0
69=SP B Port 1Typically the customer will have 2 FC switches (failover).
Switch1 will have 2 VNX connections:
alias: SPA_P0 50:06:01:60:3e:a0:5b:f8
alias: SPB_P0 50:06:01:68:3e:a0:5b:f8
Switch2 will have 2 other failover VNX connections
alias: SPA_P1 50:06:01:61:3e:a0:5b:f8
alias: SPB_P1 50:06:01:69:3e:a0:5b:f8
DS300B alias

DS300B alias

2. create Zones

I’m using Single-HBA zoning, so each zone created has only one server-HBA (initiator) in the zone together with the aliases created for all of the target storage devices.
DS300B Zoning

DS300B Zoning

3. create Zone Config

Add all zones to the Zone Config which will later be saved and applied.

DS300B Zone Config

DS300B Zone Config

4. Save Config
5. Enable Config

Useful commands:

#tstimezone –interactive
#date “mmddHHMMyy”
#switchname “name”
#chassisname “name”
#licenseadd “key”
#zone –validate “zone1”

Inactive new paths to ESX4.1 hosts on EMC VNX

Last week I was performing an EMC² VNX5100 FC Block implementation and ran into a “hidden feature”.

The environment was build with 2 VNX5100 storage systems connected to 2 redundant DS-300B FC switches and 5 ESX 4.1 hosts with dual HBA’s connect to both FC switches.

3 FC cables were missing to 3 ESX hosts, so they were temporarily connected with a single FC connection.

Zoning was setup and on Unisphere I configured a storage group containing the ESX hosts and LUN’s.
On vSphere I performed the necessary rescan for storage and configured the VMFS volumes.

The 3 ESX hosts with the single connection of course was showing less paths to the storage.
Then the Fibre cables were delivered and connected.
We modified the zoning to reflect the WWN’s and were expecting that after a rescan on ESX this should have been ok, but in fact the NEW PATHS to the storage were not visible on ESX.

Hmmmm…. what’s going on ??????

I doublechecked the zoning, which was ok and Unisphere was also ok.

Anyway, I went into the “Connectivity Status” and the strange thing was the fact that the 3 hosts were not showing up as part of a storage group at all, while before I already added them and they were already using some LUN’s.
So this was very weird to see.

Then I went into Storage Groups and enabled “Engineering Mode” which will show you some hidden system storage groups like “~filestorage”, “~management” and “~physical”.
I looked into my regular Storage Group for the ESX hosts and the hosts are added.

So then I opened the “~management” Storage Group and noticed that the 3 ESX hosts were ALSO members of this group and I could not remove them.

This is a problem, because of course a host can only be member of 1 Storage Group at the time !!!!

The first thing that came to mind was to remove the ESX host from my regular Storage Group, but this would have been disruptive for the production-ESX hosts, so we had to find another option…

Still in “Engineering Mode”, to overcome the problem, you need to go into your regular Storage Group, open the “Hosts” tab and in “Engineering Mode” you will have an extra button “Advanced“…  click “Advanced”
You will then get a list of the hosts and “Used Host Connection Paths”.
I noticed that my 3 hosts were showing the missing paths, but were nog selected.
After I selected the missing paths on the hosts and saved the Storage Group they were automatically removed from the “~management” storage group.

I only had to perform a last Storage Rescan on the VMware Hosts in vCenter and all paths were ok.