HPE BladeSystem Reviews

The HPE BladeSystem is a universal platform for server infrastructure. It is easy to manage and to connect to your other infrastructure, fiber channel network, and so on.

It's mainly focused on management and reliability. It's a fairly reliable platform, almost no outages. It works perfectly.

It could be improved in terms of management, in terms of uptime. When you do the firmware upgrades, it's not acceptable; we have downtime issues. It’s not good with that, but it's getting better and better.

We have been working with HPE BladeSystem C7000 since 2007. Until recently, the firmware updates on the connectivity modules (FC and Ethernet) and Virtual Connect could not be done without downtime. For an enterprise system, this is not acceptable. It is only since last year that we did the first online upgrades without any downtime.

It's very stable; just minor issues; no big issues.

You scale within the enclosure. You get 16 servers and then you can buy extra enclosures. It's scalable.

Technical support is OK. I'm not directly working with support myself, just indirectly. But from what I hear from the engineers, it's OK.

We previously used other HPE servers, just the old ProLiant servers and other lines. We converted to BladeSystems and these products.

Initial setup is quite complex. You have to think before you start.

We chose this solution quite a long time ago. I don’t remember what else we considered. We chose HPE because we were already an HPE customer.

Invest in preparation. The HPE BladeSystems are being succeeded by the Synergy systems, announced last year. That's the successor, so look at that.

We like the wire once approach. We've had to do very little maintenance on the backend connectivity components. We really have been able to basically leave our fiber and Ethernet connections alone. I think in our ten years on two BladeSystem enclosures, we've done one infield upgrade, to raise our capacity limits on Ethernet and fiber channel. But really, the same enclosures have been doing fantastic for that entire time.

We had a lot of work to do from a management perspective as an administrator to ensure that we had the correct firmware and driver levels, and things like that, between servers. HPE fixed that when they introduced the service pack for ProLiant. OneView has expanded the idea of server profiles and added some additional management ease - including the ability to define firmware levels for blades - the way that a Cisco UCS blade does.

We've been a BladeSystem customer for over 10 years.

The ecosystem of software the comes with HPE blade has been lacking. HPE's Rapid Deployment Pack (was rebranded Altiris when I used it) is clunky solution for imaging blades and it may have been replaced. Sometimes the NPIV makes it difficult to find WWN in the fibre channel SAN when bringing a new blade online.

Early on, I would say that it wasn't stable. Today, however they test all of the firmware and drivers together now, so when we do a deployment or upgrade for firmware, that service pack has been integration tested across all of the different generations and hardware components, so we've gotten a lot more stability out of the solution since they made that change, and that's something I will say across multiple different solutions. Sometimes if they identify a problem like that as a partner, they tend to try to correct it. It may take them some time to get it corrected, but we went through the same sort of thing with StoreOnce ... Early on we had some issues with the software releases and there were some inefficiencies that in a couple generations they had it worked out and they listened to customer feedback. They try to integrate that into their solutions. That's one of the things that keeps us as an HPE customer, because we do feel that communication and feedback gets heard.

Multiple enclosures at scale can introduce challenges. That's one of the downsides of a ten year old technology. HPE is working on new solutions to really hit scale issues that they have encountered with BladeSystem, but it works fine for us. We are a small to medium sized enterprise, so we've not encountered any major issues with scalability where we are at, but we are running one enclosure per sight, so we're not really trying to do scale things.

7/10 - it can be hit or miss. We get better luck with our premium support levels. We have a named TAM for some of our systems, that works out well. Escalation managers are always good. There is good technical talent, it's just sometimes hidden by first level support. That can be difficult and frustrating at times, but over ten years working with them, I would say today it's probably a little better than when I first started. Actually, I would say it's probably improved a good bit since I first started working with them, but it's still got some room to go.

BladeSystem setup was one of my first projects when I joined the company, there was a learning curve to it. It really actually was fairly straightforward except we were trying to do everything boot from SAN, and that added some complexity and learning curve. The actual hardware solution - it was fairly straightforward and made a lot of sense. Blades were brand new back then too, so it was a foreign concept, but it wasn't terrible. I would say maybe a a six out of ten or something initially, but if I had to do it now, it's probably a lot easier.

It really gets the job done well. It's dense compute. We replaced an entire rack of equipment with ten use, twelve use, I can't remember how many use it is, but with a single BladeSystem. It has been a long-sustaining enclosure. We have field upgraded the interconnects twice to add new capabilities as technology changed and those were service affecting but not very difficult. I have worked with Cisco UCS Manager and both have similar concepts. HPE Synergy is also due to launch soon, and althought HPE has a commitment to BladeSystem for the next few years, Synergy is really their next generation of blade technology.

It really gets the job done. The management is easy, it's been easy from an upgrade perspective and maintenance. So, it works really well for us.

The simplified design of the blades ensures a effeccient and repair including a replace in the event of a failure.

Improves capacity and aids organizations in adapting to their growth.

The profiles needs a complete rework, the way the profiles and networks are assigned to the blades isn’t user friendly and is often a hindrance

We've had no issues with the performance.

Excellent scalability, good flexibility from both a vertical and horizontal perspective.

 Consider implementing HP One-View into your environment to assist in addressing the complicated admin in some areas. Implement where possible a Use Standard to the chassis level, i.e. 1 chassis for hyper-v, 1 for SQL, etc... Segregate the networking set to the network where possible and avoid using the chassis profiles.

I think the c7000 platinum enclosure is really a great value for the money, especially for centralized data center management. It provides us with an advanced server management system that makes things work faster for our NOC and server administrator.

Firmware updates for system components are also valuable.

I would like to see improvement regarding scalability and deployment in the area of support. I simply mean seeing other fantastic features like built-in humidity or temperature sensors (especially for the market outside Europe and US). With technical support, I'm wondering if there could be some sort of seminar or webinar organised for users and prospective clients that may want to take a bake off on HPE BladeSystem.

The HPE BladeSystem is easy to manage. It is a fantastic product that datacenter managers, cloud services providers, and server room operators should embrace because it has all you could ever think of.

Probably the most value there is the compactness and the density you can get into the data center.

It would be nice if it could have better monitoring in terms of iLO and things like that. Some more interoperability between the different generations of blades.

Most recently, with some of our blades in our VDI environment we've had some stability issues. I don't know if that's a problem of HP or if it's a problem with HP with VMware and the VM-HP virtual connect on the backend of the Blade Chassis, but we have had issues in that environment.

We used VMware as a hypervisor on top of our BladeSystems, on our Blade Chassis, and on our blade server. It made it much more scalable for us to be able to spin up new architecture fairly quickly. What we've done is rolled back or moved, relocated some blades from one Chassis to another that didn't have virtual connect.

It's just as they always have been. They've been pretty solid in terms of technical support.

Most everything we had prior to building up our new data center was single server DL380s or 360 single servers. It would take too long to provision a DL380 to give people the access to a single server to be able to then do their development or build a new production environment.

We have a gentleman who's primarily our BladeSystem expert.

Make sure that if you're using it as a hypervisor, that you're doing all the pre-work on setup such as knowing the interoperability between different product statures that you're going to run on it.

I think the thing about BladeSystems is the quality of the engineering that goes into them. They have a long history of being valuable and viable products that are out there. Customers trust them.

It's interesting because I think, as infrastructure becomes more and more invisible and application becomes more and more important to the business, just not having to worry about that infrastructure is the value on a business level.

Customers are always looking for more performance, just trying to get more out of them. I don't know whether they have a blade or whether, but it's a rackmount server, and they're just trying to get more horsepower out of them. Continue to make them more scalable inside the box in terms of CPU, memory and I/O etc. It's just, customers are always looking for more density.

From the folks that I know that are using them, typically you're not seeing stuff at the infrastructure level. You may have some brittle pieces of the application and the integration, but the platform themselves are solid.

Architecturally, there's no problem adding more of those as the environment needs to scale.

I'm not that hands-on too much on the setup piece.

I think it's the advice that I'd give to anybody that would ask for it. Start with your application, find out what the requirements are, think about what it's going to need in the future, then begin architecting your solution there.

Valuable features include the speed at which you can deploy a system, the server profiles, and the ability to move those profiles around at will. We can also swap out hardware as needed, which is probably the saving grace for it.

Most of the issues that I have found have been addressed inside OneView. OneView 3.0 will allows us to do live migration of the Virtual Connect domains, so there isn’t anything big to improve at the moment.

We have used the product since it first came out and since generation one. We went from the P-class to the C-class because the P-class was not very good, and we haven’t had many issues since. Along the way, we went from pass-throughs to virtual-connects to FlexLOM and more, and everything has worked fine.

There were some issues with Virtual Connect not recognizing certain devices in OneView, but they have been addressed. Most issues are addressed quickly. We also use OneView to create the profiles on most of the new BladeServers that are Gen8 and higher. There are no issues with that.

It’s pretty scalable. I came from an environment that had 60 Blade enclosures. We went to Virtual Connect Enterprise for manageability, because the domains were locked to four when you did a stacking cable. There are some limits inside the Virtual Connect Enterprise Manager, but we couldn't find them.

HP support is very good. I've never had an issue with it. HP stands behind their product so they work hard to fix issues.

To pick a solution, we generally create a matrix and then fill in what we want out of the product. We pump in vendors and choose whoever meets the targets that we set. I would also note that the migration from rack mount to BladeSystem is not a one-to-one, so read the manual.

The small footprint that they have, the reliability and the ease of use are key factors. We've been using them for probably 12 years now. I've been very happy with the product. We have them in a data center, so we have to pay for the power and space. That's two key factors, as far as pricing going, as far as why we wanted a small footprint.

We don't have to hire as many IT people because of the ease of use. That's key. With the smaller footprint, less power consumption, that saves money, and that's the bottom line.

I guess with its ease of use, as far as the configuration of the virtual switches, and things like that all need work. We just upgraded to 10 Gig on a couple of them, and the learning curve for me was a little tough on that.

It's easy. If we need to add servers, it's very simple. We don't add and subtract a lot of servers. Our environment is pretty stable. I'm not looking for hyper-quick deployment of servers and things like that. I will look into being able to get into where if a server fails, that profile can quickly go over to another profile on another server. That would be nice to have that feature.

We looked at Dell and Cisco, and we actually just re-evaluated them again last year. We decided to stick with the HP because we were happy with the product. I guess because I was the decision maker, and I've been happy with HP. Unless there was a real business reason to switch, and there wasn't, so we stuck with HP.

If you're considering it, you want to try out all the 3 big players. Then kind of just go with what feels right for you. I've tried out all three of them, and I've been happy, and the HP is the best. You just got to try it out and see what you think.