Access Logix

Access Logix

Access Logix, often referred to as ‘LUN Masking’, is the Clariion term for:
1. Assigning LUNs to a particular Host
2. Making sure that hosts cannot see every LUN in the Clariion

Let’s talk about making sure that every host cannot see every LUN in the Clariion first.

Access Logix is an enabler on the Clariion that allows hosts to connect to the Clariion, but not have the ability to just go out and take ownership of every LUN. Think of this situation. You have ten Window’s Hosts attached to the Clariion, five Solaris Hosts, eight HP Hosts, etc… If all of the hosts were attached to the Clariion (zoning), and there was no such thing as Access Logix, every host could potentially see every LUN after a rescan or 17 reboots by Window’s. Probably not a good thing to have more than one host writing to a LUN at a time, let alone different Operating Systems writing to the same LUNs.


Now, in order for a host to see a LUN, a few things must be done first in Navisphere.
1. For a Host, a Storage Group must be created. In the illustration above, the ‘Storage Group’ is like a bucket.
2. We have to Connect the host to the Storage Group
3. Finally, we have to add the LUNs to the Host’s Storage Group we want the host to see.

From the illustration above, let’s start with the Windows Host on the far left side. We created a Storage Group for the Windows Host. You can name the Storage Group whatever you want in Navisphere. It would make sense to name the Storage Group the same as the Host name. Second, we connected the host to the Storage Group. Finally, we added LUNs to the Storage Group. Now, the host has the ability to see the LUNs, after a rescan, or a reboot.

However, in the Storage Group, when the LUNs are added to the Storage Group, there is a column on the bottom right-side of the Storage Group window that is labeled Host ID. You will notice that as the LUNs are placed into the Storage Group, Navisphere gives each LUN a Host ID number. The host ID number starts at 0, and continues to 255. We can place up to 256 LUNs into a Storage Group. The reason for this, is that the Host has no idea that the LUN is on a Clariion. The host believes that the LUN is a Local Disk. For the host, this is fine. In Windows, the host is going to rescan, and pick up the LUNs as the next available disk. In the example above, the Windows Host picks up LUNs 6 and 23, but to the host, after a rescan/reboot, the host is going to see the LUNs as Disk 4 and Disk 5, which we can now initialize, add a drive letter, format, create the partition, and make the LUN visible through the host.

In the case of the Solaris Host’s Storage Group, when we added the LUNs to the Storage Group, we changed LUN 9s host id to 9, and LUN 15s host id to 15. This allows the Solaris host to see the Clariion LUN 9 as c_t_d 9, and LUN 15 as c_t_d 15. If we hadn’t changed the Host ID number for the LUNs however, Navisphere would have assigned LUN 9 with the Host ID of 0, and LUN 15 with the Host ID of 1. Then the host would see LUN 9 as c_t_d 0 and LUN 15 as c_t_d 15.

The last drawing is an example of a Clustered environment. The blue server is the Active Node of the cluster, and the orange server is the Standby/Passive Node of the cluster. In this example, we created a Storage Group in Navisphere for each host in the Cluster. Into the Active Node Storage Group, we place LUN 8. LUN 8 also went into the Passive Node Storage Group. A LUN can belong to multiple storage groups. The reason for this, is if we only placed LUN 8 in to the Active Node Storage Group, not into the Passive Node Storage Group, and the Cluster failed over to the Passive Node for some reason, there would be no LUN to see. A host can only see what is in it’s storage group. That is why LUN 8 is in both Storage Groups.

Now, if this is not a Clustered Environment, this brings up another problem. The Clariion does not limit who has access, or read/write privileges to a LUN. When a LUN is assigned to a Storage Group, the LUN belongs to the host. If we assign a LUN out to two hosts, with no Cluster setup, we are giving simultaneous access of a LUN to two different servers. This means that each server would assume ownership of the LUN, and constantly be overwriting each other’s data.

We also added LUN 73 to the Active Node Storage Group, and LUN 74 to the Passive Node Storage Group. This allows each server to see LUN 8 for failover purposes, but LUN 73 only belonging to the Active Node Host, and LUN 74 belonging to the Passive Node Host. If the cluster fails over to the Passive Node, the Passive Node will see LUN 8 and LUN 74, not LUN 73 because it is not in the Storage Group.

Notice that LUN 28 is in the Clariion, but not assigned to anyone at the time. No host has the ability to access LUN 28.