UPMC is the largest employer in Western Pennsylvania. The healthcare system is the first of its kind in the region to have an integrated provider-payor approach to healthcare. IBM Power Systems plays a key role in supporting both the provider, Electronic Health Record (EHR), and payor (UPMC Health Plan) lines of business. Patient care necessitates 24/7 availability, and PowerVM along with technologies such as PowerHA and LPM (Live Partition Mobility) have allowed us to minimalize downtimes. Additionally, we are in the process of implementing VIOS SSP (Shared Storage Pools) and PowerVC to automate deployment and provisioning of new workloads.
My name is Valerie Smith and I have been working with IBM’s Power platform for nearly 7 years now. Shortly after graduating college, I started out working as an IBM i administrator. UPMC has always had a large Power footprint, IBM i comprising a portion of that.
About one year ago, my job responsibilities shifted and I started working as a System Engineer for the UNIX (AIX, Linux on Power, and RHEL) team. Given my previous experience, the transition was smooth. While still working on the hardware and architecture I knew and loved, I had the opportunity to dig even deeper into the technology.
Due to my aforementioned background, I was invited to work on the IBM Systems ESP (Early Support Program) evaluating the new POWER9 system. In the simplest terms, think of it as being a beta tester. Up until 6 months ago, I had never heard of this program, but it has become such a valuable experience for me in many ways since it allowed me to,
Our trial machine was a 2-node Power Systems 980. The project kickoff was a 3-day meeting/training at IBM’s Austin campus packed with sessions going into great detail of every new and need-to-know feature of the new POWER9 system. Many sessions were led by the very developer or engineer who led the project! Although there was a ton of information to be covered, the environment was relaxed and laid-back. We were given plenty of opportunity to interact with the IBM team and ask questions pertaining to our environments.
In the field of technology, there is always something new to learn. When I was working with IBM i, I would always enjoy reading the blogs of those who were trying out the latest releases of TR (Technology Refresh) code or software, such as IBM ACS (Access Client Solutions). I found it most helpful when those individuals would post little things to look out for from a customer perspective. You know, the kind of stuff that isn’t found in the IBM KnowledgeCenter. Well, imagine my excitement when I realized I get to see what it’s like to be that person! Here, I’ll share some details of my personal experiences.
About a month and a half after returning home, the new hardware arrived on site at my company. The IBM team scheduled a 3-day period shortly after to come onsite and actually install the machine. The 980 was a little longer in size than the 880 (and as I have heard the 950 even longer!) which can cause installation issues for customers with non-IBM racks. It truly is the little things! The Power-based HMC we ordered was the first of its kind at my company. A POWER9 server can only be managed by an HMC code at a minimum of V9 R1 910, so we were wondering if the HMC would arrive onsite with V8 or V9 code. It arrived with V9 so we only had to do a small update, rather than an upgrade, to R1 920. I noticed when the system rebooted after the update, that the console would display a black screen, for what seemed like an endless amount of time, with no indication of progress. I was able to pass that feedback to the IBM team, who could then get that information in front of the HMC development team for addition to a future update. Throughout my years working with IBM, it has stood out to me that IBM always welcomes user input with open arms. This ESP Program embodies that! Yes, we as a company get to try out the new hardware and code early, but what I would argue is of equal importance, is that we, on behalf of our Power user community, get to provide that real-world feedback to IBM. They asked us to share our feedback at every step of the way.
After the 980 was racked and HMC ready to go, it was time to power on the server for the first time! Picture this - I was standing behind a rack in our datacenter in front of a keyboard/monitor/mouse set up on a crash cart, driving this effort, while a semi-circle of IBMers were all standing behind me, observing and commenting on every little detail. It was a blast and felt exciting! We found one small issue that wasn’t able to be immediately resolved. When attempting to test an FSP failover, the HMC reported that it was unsuccessful, yet the Primary and Secondary FSP IPs reported as swapped. This issue is promised to be fixed in later levels of code. For me, it was cool to see all the work that goes into making the Power servers as reliable as I have always seen them to be!
After the IBM team left, we continued on with the work of planning and configuring network and storage adapter layout. When we had questions on which slots allow which types of adapters, a quick email got us the answer we needed. My personal favorite feature of the 980 so far is the addition of the NVMe drives to the CEC that we will be able to use for VIOS boot disks. I’m also looking forward to trying out VIOS 3.1. One thing that surprised me at first but I can understand now, was the multiple updates to server firmware and HMC code. IBM put the pre-GA code for us on a file sharing site, and I downloaded it and applied to my machine. This was a different experience for me, because with HMC and firmware updates, I had grown accustomed to the HMC being able to go out directly to IBM and query for the latest updates to just download and apply in one step. There is value in experiences such as these, since it always helps to be familiar with more than one way to do something.
Last but not least, I especially am thankful for the myriad of personal relationships I was able to form with IBM specialists. I had the opportunity to work with everyone from the talented men and women who installed the machine - IBM SSRs, to the subject matter experts in Product Engineering and Product Manager specialties. We continued through the ESP program by holding weekly status calls. The same people I met in Austin and that came onsite for the install took time out of their schedules to meet with us each week and answer any questions. The ESP program will wrap up in December, but I know the benefits of being part of the this will continue on infinitely after.
Given my experience, I would encourage others to participate in future ESP program. Not only will it be a rewarding for you, you will help improve the end product for you and other Power customers!
Have questions for the PowerVM team or want to learn more? Follow our discussion group on LinkedIn IBM PowerVM or IBM Community Discussions