What is our primary use case?
I've been working with OneView off-and-on for about five and a half years, first on the operations side, mostly managing hardware and maintaining it. More recently, I've been spending more time on the engineering side, making sure that we're building stuff, seeing what we can do as far as automating, and getting all of our profiles more generalized so that there are fewer questions about what a particular piece of hardware has.
How has it helped my organization?
From the engineering side, we're seeing some nice benefits from just being able to pull everything in and configure it without having to spend time running down to the data center, spend time rebooting anything. We can make our configuration changes and then, a month later, when everything goes through reboots, it just picks it up. So we can plan this stuff out a lot more carefully and more cleanly.
What is most valuable?
From my early days, the big value was in hardware replacement and repair. We could literally swap out a piece of hardware, slide one back into the chassis, and immediately - about three reboots later - it was identical. We didn't have to worry about configuring it, we didn't have to spend any time getting anything into place. It came back and it healed the environment almost immediately.
What needs improvement?
The area that I'm currently looking to explore is API-based automation. Thus far, I haven't gotten very far in it, so I can't really say where we are at, looking for that. Other than that, at the moment, we don't have a big need for more features. We need the time to really dig in to it and expand our usage of it.
For how long have I used the solution?
More than five years.
What do I think about the stability of the solution?
We had one issue a couple of years ago with a slightly older version, where it became incredibly unstable. It took us several days of working with the techs to get it back into a good state, and it required multiple escalations.
Outside of that, it's been very good overall. The main problem that we run into, as far as stability goes, is when something loses its profile. Sometimes it requires jumping through a number of "hoops" to really get it back.
What do I think about the scalability of the solution?
Broadly speaking, it is scalable.
The main issue we had - I think we were on version 2 at the time, and version 3 had just come out - was actually a scalability problem. We had some stuff that fell out of the environment due to a data center move, and it was taking so long to time-out on the old stuff that it would reboot itself, repeatedly, without ever becoming functional.
Otherwise, we've added hundreds upon hundreds of systems into it, dozens of new profiles, clone profiles, everything. And it has performed to the performance level we expect.
How is customer service and technical support?
The only time we had to call, it was a very weird issue, so I think it's a little difficult for us to comment on tech support. When we have had minor issues, they've been resolved very quickly, their response has been very good. The bigger stuff has required multiple escalations. We did wind up getting the right people on the phone, it just took a little while.
How was the initial setup?
The setup is very straightforward. We loaded the virtual appliance, pointed at the old one, pulled the database in, and we were already running.
What was our ROI?
The time that we have saved on hardware configuration and reconfiguration, both on the operations and engineering sides, would definitely pay for the product.
The very simple fact if we have to buy another chassis, or we're configuring a chassis, we can take an existing profile, clone it out, and put it on the new hardware and it just reconfigures all the hardware for us, that saves us time. Instead of manually reconfiguring the chassis, and 16 blades, and getting all the networking done, and then getting all the Fibre Channel done. We just hook it all up and we spend a few minutes doing the networking and then we're done.
What other advice do I have?
At this point, I'd rate it about an eight out of 10. It's doing the job that we need it to, it's doing that job well. Sometimes the interface can be a little confusing, sometimes the error messages can be very cryptic, as far as what's actually going on. But on the whole, the product does what it sets out to do and does it well.