FlexPod XCS Reviews

The most valuable feature is that it's an end-to-end solution for which we have a single support structure. We can tag it in the event we have something that crosses between silos.

We like that we've got the validated design that we can use quickly to start it. We have something that we know that will work. We also like the fact that the whole end-to-end design has been tested. And we know that if we have a problem, we're in good company possibly. That's a big help for us.

The benefits to my organization are faster time for deployment, faster time to problem resolution; a stable and reliable environment.

I think it still seems that there is isn’t end-to-end automation. I can't say, "I want to build a host," and say, "Provision the blade. Provision the storage. Provision the SAN." All those things in one swipe would be nice. I think that's what everyone wants: push the button, make it work. The big thing is that we'd like to be able to have it provisioned end to end. The biggest weakness is on the SAN side and the fiber channel side is a mess. It isn't very clear at times. That part is the most complex. If we can avoid it, we don't do it.

It's been a very stable environment for us, absolutely. We've only had very few instances and each of those have been resolved very quickly. It's been a fantastic change for us.

It just works, scalability-wise. We've been able to continue to grow the environment. We've tripled in size since initial deployment. We've not really changed the overall structure of the deployment. Scaling up was very easy.

Technical support has been very good. We haven't had to use it much, which is good too. We've had once incident where we had to really get all three legs involved – the compute, the storage and the network side involved – and it worked well. It was a very good experience to see from end to end. The problem looked like it was storage, but could've been on the SAN side. We worked through that and we also had to go back through the host. We had all three parties involved. They worked well together.

Initial setup was very easy. We looked at the validated design. We tweaked it a little bit to fit our environment specifically. We deployed from there. It was a day or so at best, more because we have our own little processes we have to go through.

ROI is difficult to gauge because things have changed so much in the time since we started. We have moved from a virtual environment of roughly 30, 40, 50 percent to 70, 90 percent now. I want to cut it.

We've been using HP for a long time.

For storage, we were using EMC before. We had been using their platform for a number of years. We were in for a forklift upgrade. The next upgrade would've been a complete forklift upgrade. We evaluated Hitachi, NetApp, and EMC. We talked with Dell as well, but we really narrowed it down to Hitachi, EMC, or NetApp. With EMC, because it was a forklift upgrade, we were really upset that it wasn't something that we could just add on to what we had or change out the parts. We had a major conversion, so that put them at a disadvantage right away. In talking with NetApp, the big thing for us was getting rid of the forklift upgrade; going to a system where we could evolve from year to year to year. We're three or four years in on this now; so far that's been the case. We have done controller replacements and upgrades and everything else, without serious service interruptions, so that's been a big deal for us.

You need a partner that you can trust and a partner you can actually move forward with. You need to understand what the full lifecycle of the product will be. You need to understand where you're going to go next. That was the big thing for us. When we changed, it was a major change. It was a lot of downtime, a lot of frustration for a lot of people. That's why something like NetApp, where I know that I have the ability to take out the controller, grab a new controller with cluster data ONTAP; you've got the scale out ability as well. That's been really great for us as well. As I’ve mentioned, we've tripled the amount of storage, we've tripled the amount of compute, added additional controllers, with little interruption to users. That's huge.

A couple of the valuable features are the ease of data management and data portability. That would probably be the crux of it. That came into effect with migration from or upgrading an existing cluster to another cluster; it is quite easy to port across using similar technologies, etc.

Obviously, it has the high availability features; you've got storage failover, or takeover, to upgrade a node non-destructively. Those would be the main valuable features.

The main thing is providing high availability to internal business customers, whether it be just day-to-day operations, HA and, again, non-destructive maintenance.

Specifically, with ONTAP, I would like to see it support more snapshots for retention; greater than 255 snapshots. There are competitors out there offering far more than the 255 snapshots per volume; Nimble, actually, being one of them. Their current marketing blurb is 100,000 snapshots. I need to find out some more details on that. I’d like to see very long-term data protection, an increase in the number of snapshots.

I haven’t had any issues in the day-to-day operations. We have had a few stability issues with some major bugs with it, with ONTAP. We had a particularly nasty bug with one of the QLogic CNA cards, which actually took, probably, seven months to resolve. Eventually it was resolved. It was actually a new bug that was still under investigation with NetApp but that was eventually solved. That's been the only really major glitch recently.

Overall, it's been a very stable platform.

I used technical support for that bug, in conjunction with our SAN managing that process.

Technical support was very good. Obviously, we've used on other occasions as well when we've had difficulties and they usually have been resolved quite quickly.

We've been a NetApp customer probably since 2005, not so much in the FlexPod but actually from the initial FAS arrays. Because we already had the FAS arrays or the FAS systems, it was a natural progression, really. We were looking to refresh or move away from HP for our compute. With FlexPod, we already had the NetApp storage and it was sort of a natural progression. The price point was pretty good.

Initial setup was quite straightforward. I guess our first FlexPod implementation was with one of our data centers. It was a sizable deployment but quite straightforward.

At the time, there weren’t any other vendors on our short list. I guess we were looking at EMC in the early days but I don't remember what the reasoning was. We eventually decided on FlexPod.

When I choose a vendor, cost and ROI are the most important criteria that I look for.

Follow the best practice guides and you can't go wrong.

If you've already got a storage compute environment, a good analysis of what your existing environment is like, what your workload's like, will assist where you place data, where you place your volumes on the FAS system. You generally have a good understanding of what your existing workload is, so you can obviously scale and design your storage layout with FlexPod correctly or most efficiently.

I guess I'm a bit biased in my opinion but I've worked with it for probably seven, eight years. I’ve never had an issue with it and everything that I need to do with it, I've been able to do pretty much.

The most valuable features are the validation by NetApp and Cisco, that it is preconfigured and preconfirmed, the flexibility and the ease of deployment.

It can spin up sites quicker and faster and in production, so it cuts the lead times and results in more production.

It's always the same: upgrade and expandability into the future; maybe a little more forethought on that versus having to have outages when you're going over to the next feature. For instance, a smooth transition, because changing from 7-mode to cluster-mode is a little clunky and then you add on to that. But, I know, that's technology.

They've done an excellent job going into cluster with version 9 and the later versions of 8. Everything's more GUI, so you have a choice of doing command line or doing GUI, whichever works better for you at the time. I thought that was an excellent change for them versus just being command line.

Nothing's perfect. Everything can improve. Just because I haven't thought of it or haven't hit it, it doesn't mean it's not out there.

It's pretty stable. It depends on bugs that are found. For the most part, it's been very stable for us.

It’s very scalable. We've been able to connect different sites and add them on where we've need to grow and then shrink down and move things; I like that a lot.

We just used technical support to do firmware upgrade. It was very favorable. It was a professional, concise, quick, to-the-point answer. I enjoyed actually engaging them.

I was a recommender of FlexPod. The decision to go with it was made by management.

I've previously used the Vblock solution. I've used the HP solutions as well. EMC is overly complicated, disparate systems kind of lopped together, and I don't like their management interface that much. HP has a pretty strong solution as well. The FlexPod is a bit more integrated, consolidated and easier to deploy. Between the two, I would choose NetApp. If I didn't have a choice, HP would probably be my second.

The HP solutions are a little complex. Support is not as swift as with the NetApp FlexPod solution. The advantages of HP are similar to NetApp: it's one-stop-shop, one SKU, one deployment, a prevalidated system.

The Vblock is okay with EMC on it. Having that solution where you can get one SKU and scale it out, if you choose to go EMC, is good, but that's about it .

We do deployments all the time. I was involved in the last one. We deployed FlexPod for one of our locations. The documentation and validation for it was very simple and easy to use, as compared to some other products that I've deployed.

I think it was easier, in addition to experiencing it before, because of the way the instructions come in, the support, the setup of how you actually physically assemble and connect the components, and the ease of management definitely put it ahead of other solutions.

FlexPod is definitely easy to deploy and go with. If I had to recommend it, I would definitely recommend the NetApp FlexPod solution.

The most important criteria for me when selecting a vendor to work with is that their product works. That's the most important thing. Then second is customer service and getting to solution. I hate a lot of side talk, empty promises – nothing becomes of it – just to get the sale. Really, make sure the product works and then you get the support that you need and not chatter.

It's easily integrated. We have WFA and Cisco UCSD; that integration works quite seamlessly. So far, for self-provisioning and for our internal cloud, for the private cloud, that integration works nicely. That's a cool thing.

The integration and all of the levels of automation, everything is there. The support is the first thing that needs to be improved. The second thing is that because there are so many products out there, they need a common management tool. I don't think they have that particular management tool, a centralized tool. I can see into WFA and so on, but that is still in the initial phases. I would like to see more on that automation.

I have been using for the past year.

It is stable; it's semi-stable. Because cloud IT itself is new, standards and procedures are likely not universally defined. It is semi-stable.

That's the challenge. Scalability is the challenge; running out of network ports, running out of storage. If you have very unpredictable workloads, then scaling is a big challenge.

Support is something that has gone a little bit downhill in the past two years. They don’t have a complete devops model. Looking over the past eight years I have been using NetApp, over the past two years, it's not been that great, compared what is used to be. Response time is the challenge.

The company has been using FlexPod for a while, actually; three years.

For the customer, upgrades are very straightforward; nothing to worry about. You just click a few things; it's easy. The upgrades are not that much of an issue.

Go for it. One thing to be aware of is the support, but the rest is good.

When I’m selecting a vendor such as NetApp, in the business that we are in, I look for stability. When it comes to stability, reputation will matter; how long have they been in business and for how long they have proved themselves. The reputation is one of the things. Stability is the key thing.

We really like the flexibility. I love that – it doesn't happen a whole lot, but – anytime we have a blade die or even just bad RAM, I grab that service profile, throw it on another blade, and we're up and running without any issues.

We're also a pretty siloed environment. I love that we're not having to add new ports and stuff every time we add a new server; we already have everything connected. Worst case, if we're adding a new VLAN, networking adds that VLAN to the trunks to the fabric interconnects, and we're up and running without a long process. It used to be, we would have new builds run up to weeks, whereas now I roll out a new VM in 15 minutes; if I have to add new network connectivity, it might be an hour or two but that's it.

I like the user interface a whole lot. The new 3.1 release has the HTML5 interface. I finally don't have to mess with Java. I still have jump servers with specific versions of Java for FlexPod we've deployed at different times that are running different codebases. I don't have to mess with that anymore. I'm looking forward to that new equipment.

We actually run one of several private clouds within our company and it is all on FlexPod. We run it as a profit center. We're able to give all our internal clients the fastest response time of any of our server teams and the flexibility to, if you need a VM with one CPU and a couple gigs of RAM, or if you need 20 VMs with eight CPUs and 32 gigs each, it's all the same to us. We have the blades to support it all.

I haven't really come across a whole lot of areas for improvement. There are features I'd like to see in our deployment that are already available; all-flash trays, and items like that. It's there; I need to find a project that justifies getting it rolled out in our data center.

Everything works pretty well. I think they should just continue to add more features and capabilities for hybrid cloud, especially items such as cloud bursting to one of the public clouds. Specifically, they need to make sure that, for our client site, it integrates with the FedRAMP clouds; it's got to be Azure Federal or AWS Federal. It can't just be to the regular public cloud.

I haven’t yet come across any features in other solutions that I’d like to see in FlexPod. Some of the newer storage vendors have slightly easier-to-use GUI interfaces, but I weigh that compared to the control and other features I have, such as SSHing into my NetApps. It'd be nice sometimes for quick stuff, but it's not worth giving up the control I have with the NetApp filers there.

There's always something new down the road, something new that can be done, but I think it's doing as well as possible. It seems like they keep getting new features, new ideas out there. We have flash on all four of the different lines now, continuing to evolve more cloud control with the UCS Director; it keeps growing. I love it.

We've been running it six years now.

We've been running it six years now and I just ordered my gen-4 gear; it should be here at the end of the month. I've had no issues.

I do some consulting, too, and have deployed it for a bunch of clients at smaller scales. I've not had a complaint yet.

I’ve occasionally needed to use technical support; generally, just with drives. Half the time, our rep will contact us to schedule the drive shipment before we even notice the email from the alerting. We love it.

We’re absolutely happy with the technical support. NetApp has the best tech support, which I've heard is part of the interview process to get the job there. I've heard rumors that it takes multiple days and is brutal, but they're the best guys we've got. We deal with lots of large vendors, and NetApp definitely has the best support teams.

With our particular case, our previous company had spun us off without any IT staff, so we were using third-party IT and we were trying to bring IT in house. Because we were having to build our IT staff from the ground up, the flexibility and all of the things that FlexPod made easier means it is a whole lot easier to bring IT in-house. We didn't need multiple storage people, server people. We had it more integrated, and had the single company to call for any issues we had bringing that all up.

I was not involved in the initial setup. Another couple of guys on the team did it. We did the initial setup with Firefly, who was our vendor, and provided week-long training. We each got to set up several of the blades, but Firefly did the initial UCS setup. I've done other UCS implementations since then from scratch, though.

I love the setup. I think it's very simple. I'm biased; I'm a CCIE Data Center. I like it enough that I have put a lot of time into it. I took a client from all physical to all virtual in a four-day weekend, with all-new switches, new SAN, new UCS, and in four days I already had, I think, their first 10 or 15 servers moved from physical to virtual on UCS setups.

We ran through all the primary vendors, but it was all one-off; there weren’t any converged solutions at the time, six years ago, when FlexPod first came out. It was either work with HP, Dell, Hitachi, EMC, etc., or have only one place; it’s better to have everything in one place.

Definitely take advantage of all the training you can, particularly the UCS portion of the setup. It's very much a one time setup if you do it right the first time. I use the reference install guide for the VMware on FlexPod, even for installs that aren't using NetApp storage because it's so well-written. It's 175 pages but it gives you not only what to do, but why to do it and even full configurations you can copy-and-paste in to make sure that everything really is "set it and forget it". You can just continue to operate your business, serve your clients as well as possible, and not have to go back and try and fix things.

We’ve never had an incident where we set it up and it hasn't worked; there have just been little things, such as when I didn't think and was setting up the first of the two fabric interconnects and put dash-A at the end of the name. I forgot that it adds A and B anyway, so that client is UCS-CompanyName-A-A or -A-B. As long as I followed the guides, I've never had an issue.

I really like good support when I’m looking at a vendor; that’s one of the most important criteria for me. I'm big on vendors that have good training. I want a vendor that wants to support their clients, wants their clients to be better educated. I don't want a vendor that wants you to always call them and maybe bill up a support call. I want the vendor that gives you the training so you can operate the equipment best yourself; still have good support when you do have to call, but give you the information you need so you can do it yourself and operate it as well as possible.

The initial reason why we got it is because it's a one-stop shop for your compute, and the networking aspect of it. We wanted a sort of one-throat-to-choke approach; if we ever had any problems, we would get the teams onboard instead of finger-pointing back and forth. That's probably the biggest draw of FlexPod for us.

It simplifies the whole stack, and allows us to worry about other IT problems and not have to struggle with stack issues, basically.

We used to have HP C-series and kind of a hodgepodge network. We had, probably, four different vendors. Because of that, we were having a lot of problems with networking and the blades. Once we got FlexPod, all that stuff went away. It was nice.

The user interface isn't as friendly as I would like it to be. I would like to see UCS ditch Java and go with HTML5, just to make that a little simpler. Other than that, NetApp is making good changes. The tools have gotten dramatically better over the past two years, which has helped a lot. Cisco – I generally don't do too much of the networking stuff – but it's still kind of best to use the command line, so I guess they could improve the command line interface.

If there could be an integration between Cisco and NetApp, the single pane of glass works well if you can get it all to actually work and get the data that you need easily and quickly. That could definitely help.

I have not encountered any stability or scalability issues. This thing has been running great. We actually really like it. It works well for us, we run about 400 VMs off of it, split between two data centers but it's a good performer.

NetApp’s technical support is probably one of the best, in my experience. I've dealt with, probably, Microsoft being the lowest, because they suck, or sometimes, if you’re in Unix land, you have to just find your own solutions. I think NetApp’s good. They're right up there with Cisco, as far as support goes; very good experience overall.

We knew we needed to invest in a new solution because we were at capacity on our blades and we needed to move to something else. Cisco UCS at the time was just kind of up and coming. I think we really made the right decision, because their one-profile approach of propagating all the configuration data down to the blades was really nice.

I was involved in the initial setup and we had a consultant too. I was just getting into it so, for me, it was kind of complex but looking back on it, it really wasn't that bad.

We looked at HP and we looked at Dell, but we ended up going with Cisco and I'm really happy with it.

We've already had businesses come to us and say, "What do you use?" We just tell them about the whole FlexPod idea and what kind of benefits it brings: the scalability, the one-throat-to-choke kind of support. We've had a couple of companies take our advice and they were happy with it too.

In terms of homework to do before implementation, you have to look at your workloads and see how you're going to use it. In our experience, we have thrown everything but the kitchen sink at it and it actually performs really well. We've had very little problems, very little problems.

When I look at working with a vendor, I like one that takes an approach of making decisions in your interest. You can usually sort out pretty quickly whether they're there to just push gear or raw services, or if they're actually there to be a partner to help you succeed in the landscape. You can shake that out usually within a few meetings to see how they're really going to be. I’ve found that in NetApp.

It allows us to run a very lean infrastructure department. It doesn't require a whole lot of administration, so I'm able to have two sys admins for our entire organization.

The user interface could be a little bit more intuitive but overall, once you get used to the way it works, it works great.

We use SnapMirror to back everything up, so I’m pretty happy with it right now.

It has been rock solid. It has delivered everything that we've asked of it and we've had no performance issues.

I have not really needed technical support. Once we got it all set up and configured, we've only had to reach out to NetApp to replace failed drives.

We were using the IBM-branded X series, so we had NetApp storage in the past. It was reliable, it was great. We really didn't a see reason to move away from NetApp. At that time, we were building a brand new data center and got the opportunity to do a full build out. That was when we selected the FlexPod.

Initial setup seemed a little complex, and I think a lot of that came from the UCS side of the FlexPod. Obviously, it is very powerful; it’s very customizable. Once we got through all of that, once we get past that, it doesn't require a lot of messing with it or hand holding.

The UCS side was complex with the levels of all of the settings that needed to be applied in setting up your port groups, and similar items.

We selected FlexPod because, as we're building our new data center, we were looking for something that would be able to scale easily with the business, without locking us into having to do huge forklift upgrades. We just had the opportunity to have a full greenfield build. We looked at other solutions such as Vblock. FlexPod was able to meet all our needs without too many problems.

Do it. Once you get it set up and configured, it's rock solid. It scales with your business easily. It will make your life so much easier once it's in place.

It's so reliable; I have had no problems with it. It's delivered everything that we've asked of it and we haven't even moved to flash yet. I'm pretty happy with it.

As far as anything that you need to prepare beforehand, I'm not entirely sure. We got to do a full greenfield upgrade. We weren’t trying to do a rip-and-replace of our current environment.

The most important criteria when I’m looking for a vendor is reliability, both with the product and the vendor relationship; not a lot of turnover in account managers. Things like that.

The most valuable features are the integration and the ability to have support, planning and installation from a single service provider.

Integration between the UCS blade side and the NetApp side is excellent.

The ability to really leverage the 10-Gb connections between the UCS to the Cisco Nexus switch and then to our NetApp really improved performance for us and allowed us to experience a huge amount of growth with no loss of performance.

Also, we've been able to move to and implement a new ERP system, J.D. Edwards. Because of the modularity of the system, when we need more compute resources, we just buy more blades. If we need more disk, we just buy disk shelves. They integrated very easily.

It simplified our workflow.

Right now, we have no flash at all in our NetApp side, so one thing we're looking forward to is going to ONTAP 9. We're also looking forward to looking at integrating some flash shelves to see what the performance will really be. Everybody tells me it's fantastic.

We're rolling out J.D. Edwards location by location so the amount of performance we're going to need is going to grow and grow and grow. So far, there's been no problems but I like the fact that I have that growth path to put in flash and improve performance if necessary.

In a perfect world, I would also love to be able to manage everything from a single pane of glass. I think we're talking about such disparate technologies that I would understand if that is very difficult to happen. In our environment, I'm in control of Cisco UCS manager and the NetApp side but when we get to the Nexus switches, I don't even have the log-ins, our networking guy does. That's something that I don't have a problem with. He's very good and he works very well together with me. It would be nice to have control from a single pane of glass.

We have been using it for about three-and-a-half years.

We've had exactly one issue and that was a related to a hardware failure, a RAM stick, that took down one blade. It was at a SQL cluster, so the other blade just took over flawlessly. We didn't have any downtime.

The scalability is tremendous.

Before we went with FlexPod, we were still a NetApp customer. We were using Dell rack servers connected via 1-Gb links for NFS and 4-Gb fiber channel for block storage and still running VMware vSphere. Things were okay but it was time for a hardware refresh. At that time, we evaluated Dell, HP and Cisco UCS; both rack and blade servers. We pretty much eliminated HP right away. One of the reasons we decided to go with the UCS was that our NetApp reseller was very much certified with Cisco and had a good reputation. As I’ve mentioned, it would have that one source, where we could get support for everything through that reseller. It also didn't hurt that Cisco offered a fantastic deal, where they quoted a price for their blade servers almost exactly the same as what Dell wanted for their rack servers. The price is a huge factor for our company. We're a privately held company, so price is often the primary factor.

For the initial setup, I worked with a reseller. They had two awesome engineers, one from the UCS side and one from the NetApp side. They worked hand-in-hand with me and the people at MacLean-Fogg to make sure we got everything done right. That is the real key with the FlexPod. If you get all your definitions and all your profiles set up correctly in the UCS manager, then adding a blade is very simple. You put in the blade, you turn it on, you apply that profile to the new blade and you're up and running.

The big thing with a FlexPod is, you've got to get it right at the beginning and then everything from that point on is very simple.