We have a single fabric SAN connected to HDS storage and the host servers sending traffic through this are SUN servers.
Objective: To split this single fabric SAN and provide a fully redundant solution.
What we are doing: We are moving one HBA connection from each server into this fabric and moving it to another fabric called SAN B. Hence, providing a fully redundant solution.
These switches are not zoned right now but the new fabric will be.
Just wondering what issues you think we should take into consideration while performing this split and are there any common problems that occur during such an exercise that you can enlighten us about so we are prepared to handle it when it arises?
Hope I have conveyed my objective clearly.
Ok, first of all, let me say that I'm glad you're moving to a dual fabric SAN. Using two fabrics allows SAN maintenance to occur without planned down time.
Since you had a single fabric SAN, I assume you just had a single host bus adapter in each server to begin with. So the first thing you need to do is add another host bus adapter in each server. When you use two HBAs in your servers, with access to the same storage volumes, you end up with two paths to the same storage.
I understand you're using Sun servers with Solaris. Depending on the HBA type and driver version, you may need to add the second controller path to your SD.conf files for the second path. You may also have to purchase a copy of HDLM (Hitachi Dynamic Link Manager), which is the path failover and load balancing software used to automatically fail the path over in case of problems with a path or maintenance on a path. If you're using the newer Sun HBAs with the leadville driver, you do not need to edit SD.conf since the leadville driver is a kernel level driver and uses SDD.conf, which needs no editing.
I would write down the WWN of each HBA for each server so you could predefine and implement your zones before you hook up your servers to the other fabric. If you are using switches from your existing fabric, make sure they are reinitialized and there is no zone configs in place. Also, if you are using multiple switches in your second fabric, you need to make sure that each one has its own "domain ID" in place. Every switch in a fabric must have a unique domain ID.
When you create your zones, be sure to use unique alias names for EVERYTHING. This is just best practice. That way, if you ever have to ISL existing switches together from different fabrics, it will still work. (Also use unique Domain IDs for every switch even though they are in different fabrics.)
I'm not sure how large your fabrics are, but it would be best to use a "core edge" design for your environment if more than four switches are used in each fabric.
On the storage array, make sure you connect each fabric to opposite sides of the array. HDS storage arrays are divided by cache and power boundaries for high availability. You should keep your fabrics on opposite boundaries on the array. (Use Ports CL1A-x for fabric A and ports CL2A-x for fabric B.)
Hope this helps!
Editor's note: Do you agree with this expert's response? If you have more to share, post it in one of our .bphAaR2qhqA^[email protected]/searchstorage>discussion forums.
Dig Deeper on SAN technology and arrays
Related Q&A from Christopher Poelker
SAN expert Chris Poelker compares connecting a SAN with wavelength cabling and dark fiber and discusses the pros and cons of each. Continue Reading
SAN expert Chris Poelker discusses how to change the size of a LUN in a Microsoft cluster server environment. Continue Reading
Storage expert Chris Poelker outlines WWN basics in order to answer the question: "Why do HBAs in a SAN have same base?" Continue Reading