What a difference a year makes

A year ago, I had just accepted an offer to join @lynxbat at his company in Fort Worth, TX.  With the change in companies, this meant I was going to VMworld 2009 on my own dime.

Fast forward a year, @lynxbat is a vSpecialist with EMC, and I’m a Network Architect with ACADIA.  Ah, how times have changed, and all for the better.

Advertisement

Upgrading a UCS

Today I had the opportunity to upgrade a UCS from 1.2.1b to 1.3.1c.  I had watched it done before, and it is straight forward process.  However, there is one thing to pay careful attention to.

When reading the upgrade documentation, you will notice there is one piece of hardware that can be upgraded either at the beginning of the process or at the end.  This piece of hardware is the adapters in the blades.  If a firmware policy is not applied to the service profile, then the adapter can be updated at the beginning of the process.  If a firmware policy is applied, then the adapters will be upgraded as part of applying the new firmware policy.

So what is the difference??

When you upgrade the adapters at the beginning, you are given the option to activate the firmware on next boot of the blade.  This is helpful as there is no disruption of network or SAN traffic at this time.  However, updating the adapter through firmware policy is a different story.  When you apply the firmware policy to the blade, the new adapter firmware is written to the blade, and THEN activated, causing a disruption of network and SAN traffic.  When you are booting from SAN, this can cause an interesting situation.

So what to do?

There are two ways to handle adapter upgrades when using firmware policies.  Either remove the firmware policy from the blades before starting so the adapters can be upgraded first, or shutdown your servers before applying the firmware policy.  Both methods work; it’s just really a matter of preference.