NetApp ONTAP Reviews

Our primary use case for ONTAP is that our whole platform is running on ONTAP. We have all the client data that we're using on the files, in our data center.

It's hard to say how this solution has improved our organization because we've been using it for 18 years now. It works great, it really does what it should and we've been really, satisfied customers. 

Our whole platform is so big that this solution is mission critical for our company. We also have a metro cluster internally where our virtualization stuff runs on.

ONTAP has really reduced our costs because we learned that we could use our storage with fewer machines and drive down data center costs.

The most valuable features are stability and performance, and that with the cDot feature we can move data around without any interruption. Also, the hardware maintenance is really, really, easy and fast. 

We are discussing ONTAP for AI. We are having a look at it, but we haven't used it yet.

We have been using deduplication on our web volumes and have seen about 25 to 28% data reduction. That's not that much because mail storage is pre-compressed by the clients, and we do not save anything there.

Synchronous NetMirror and FlexCache features will be back again. This is really great. It will help us be more efficient but it will take some time until it really comes down and we can use it.

We hope that SSD's will be cheap enough so that we can consolidate and save reg space in the future.

Stability is really great. It's awesome. Of course, we have hardware failures, but they really work great. The failover mechanisms do what they should do and, we are very satisfied with that.

cDot only scales to 24 notes so scalability should be better, bigger, but we are one of the only customers that are facing this problem.

Their technical support is mostly very fast. Our system account manager takes really great care of that. Sometimes, parts are shipped in the middle of the night, when nobody is on site which could be improved.

Our FAS9000's do, 150,000 IOPS per head which is less than around one dollar per IOP.

I would rate this solution a ten, especially the cDot version because it really helps focus on our real problems, and storage is, not the greatest problem anymore and really works great.

My primary use case is we have both Windows and Unix and they share file systems for compiling code. The big advantage with NetApp is the ability to file Linux symlinks in the Windows environment. 

When I came on board they had NetApp and ONTAP was old and the system was getting to its end-of-life and corporate wasn't sure which way they were going to go. They couldn't quite make a decision on whether to buy a very large unit or a small unit because we were gonna become a central hub. They decided to scrap, and what to choose landed in my lap. I decided to go with a smaller NetApp that would fit the main requirements that I needed NetApp for and I use other types of storages for VMware. My volumes, that are NFS and SIFS, there's a lot of stuff that's used both on Windows and Unix so I need the ability to maintain the permissions between the two. I get better security with ONTAP and I get better control of users space requirement because I have qtrees and quotas and then I have the masking of user accounts, NIS to AD. The other thing that's a really good bonus is that ONTAP has a deprecated NIS and a lot of other vendors are deprecated NIS.

Critical applications are not as critical as like you'd normally experience because I am R&D and it is a production environment for R&D, but I have time to build a recover. I can recover hourly from snaps, everything else I recover from tape backup because my backup uses MDMP and it'll be just as fast as Snap and storage are cheaper.

Cost of storage hasn't reduced but it's more cost effective because the very specific requirements drop the ball. Especially when it comes to user account translation from Unix to Windows. ONTAP and Dell EMC are the only two real vendors that know how to do that properly.

For me and my users, the most valuable feature is the ability to mask Unix accounts to Linux accounts.

Stability's perfect because I have two nodes, I'm not overloading the nodes because it's just R&D and it's very specific lines, so it's a lot of terabytes but we're not in petabytes. For what I do it's very stable.

Scalability is perfectly fine. Right now I only have the two nodes and one shelf. I'll be able to easily upgrade additional shelves. They gave me plenty of cabling when I got the unit so all I have to do is disconnect and reconnect the cabling and that's it.

Tech support has been pretty awesome. The only thing is that 9.4 has been presenting a couple of challenges and there was one case, for example, where I didn't want Snaps. There's a command to be able to disable the scheduler, but with 9.4 that command doesn't quite work. I ended up using a workaround which tells the scheduler that it has zero snap capability on all snap jobs.

The initial installation could've been quite easy, but there was a lot of miscommunications with professional services and there are a lot of details that they didn't quite provide which caused a very complicated installation.

A lot of Windows builds have been failing simply because when they go through the file system they can't file the symbolic links that are created on the Linux file system. Now they will resolve because ONTAP supports that.

I would rate this solution a ten. It's very easy to use. What I really like about it is it incorporates the same thing as CentOS and RHEL 7 which is the Tap commands. If you have an idea of what commands you want to use, you can tap through and figure out what you need without having to go and look for the full syntax.

I have use this storage appliance, which they have on-premise. It is for taking backup off onsite data. This is a storage and NAS device.

Its typical use case is storing files and getting data, so it does its use case well.

When it accesses regular files, it does so as an NFS workload, which retrieves data quickly.

The initial UI integration could be quite a lot better.

I would like to have deduplication and SSD features in the product.

If you have huge data, you will see stability issues. Since we didn't have that data, our stability has been fine.

We previously haven't had a huge workload, so it scaled fine. Though, with our current use case, we will need something faster. E.g., if you have a huge scale, having SSD-based backup is better. 

We did not contact technical support, as the solution was working fine.

The integration and configuration are not straightforward. We had to read some documentation and receive support. A good UI may be helpful in resolving this.

We looked at various Dell EMC products at the time and considered using VMAX. We also looked at Nimble Storage. For some reason, we did not look at Pure Storage.

We chose NetApp because had some collaboration or a sales representative who approached us, then convinced us of the use case.

Based on the use case, it can be a good product.

We have just one application where we store data, and we haven't had problems with it.

We use it for the security business. We use it mostly for capacity-oriented purposes, rather than performance-oriented.

It's a major product for us. We are dealing with security stuff so the encryption features in NetApp really help, as well as the deduplication. We are able to minimize the storage hardware. The compression and deduplication have helped reduce our overall cost of storage.

The valuable features include replication, Snapmirror. That's really useful for us. Also, FlexGroup is useful as it is capacity-oriented, so we can keep extending the space.

I'm waiting for end-to-end NVMe.

It's been running for years and years and we haven't seen any issues.

Scalability is good. The NAS can expand to 24 nodes for the FAS series, and for SAN it can expand to 12 nodes.

We have had to use tech support and the response has been good.

We switched from our old solution because of the features. We went with NetApp because of the redundancy, availability, scalability, and cost.

The initial setup was easy. It took just a few commands.

We used a reseller. Our experience with them was good.

We did a PoC with a few other products.

ONTAP has been in use for over 25 years, that's one of the major advantages when compared to start-ups and other companies. It's also global, NetApp has support all over and, in case of an emergency, their response is good. When there is an issue, many people jump onto the call to try to resolve it.

NVMe over Fabrics is margin-technology at the moment, but the future will be NVMe. All storage, end-to-end, will be NVMe protocol. The speed of NVMe is good. The current existing technology is SCSI-based, one command per Cube, but with NVMe you can run 65,536 commands in each Cube, meaning 65,536 Cubes. That is really fast. In terms of NVMe over Fabrics with existing Fibre Channel infrastructure, if the hardware supports it, it should be good. As the protocol improves, there should be end-to-end support for the NVMe protocol.

We don't use this product for machine-learning, or AI, real-time analytics or other groundbreaking types of applications.

We use it to store email files, mainly JPEGs.

With ONTAP, we have peace of mind with double-parity protector RAID systems. Therefore, we can sleep well at the night, not thinking about crashing a RAID, because it's secure. We are confident with the technology. 

We use it for company websites using NetApp back-end. These are mission-critical applications.

• The capability to create no overhead Snapshots.
• The capability to restore from Snapshots: file and add volume. 
• FlexClone is a great feature. 
  
• All the features are great with ONTAP.

I would like to see more integration of the features with the CIFS and SMB Protocols. We also want integration with iSCSI. Right now, there are a lot of gap between some bleeding edge technology and the assembly protocols. Therefore, I want to see those areas improved in the future releases.

The data is always available. I haven't had a disaster using NetApp products, at least the NetApp FAS systems. It's very stable and highly protected.

You can add storage capacity on the fly with Clustered ONTAP. You can add nodes and increase the entire horsepower of ONTAP.

Technical support is hit or miss sometimes. Level 1 support is not very good. Level 2 and Level 3 (the escalation support) are very good. They are knowledgeable people, but sometimes you get some hiccup in the Level 1 support. After you pass Level 1, it is smooth sailing. There is a lot of room for improvement.

The initial setup was straightforward.

We used an in-house team for the installation, and it went well.

In the past, we have used a reseller. Our experience with them was also good.

It helps us save money using compression and deduplication. No overhead snapshots and FlexClone (with no additional space) help us save space. We have been able to save about 50 percent of our space using ONTAP.

We did not consider anyone else besides NetApp.

If you are looking for something simple to manage, but an advanced storage array, NetApp is the way to go.

Our primary use case is for serving the file server files. File storage is for the member offices and accessing that data.

From the capacity perspective, it has helped us because of the thin volumes, deduplication, and compression features.

ONTAP provides the ability to dole out Qtree shares as needed. Team volumes is a benefit.

The cloud features, the ability to go into the cloud would be a feature I'd like to see improvements on.

They could also improve on their ability to back up storage.

It's stable. It has good monitoring tools to help monitor activity and take action if necessary.

The scalability is good. 

We do look at tech guides and the Knowledge Base articles and then engage support. The account teams are always available.

We were using Dell EMC. We did an initial evaluation with the reseller consultants, based on that we went forward. We wanted to diversify everything so that we could get more benefits from the features.

The initial setup was good, flexible, forthcoming, and straightforward. 

We used CopperTech. They were good to work with and very responsive. They look at every scenario.

The deduplication and compression features have helped us save storage costs.

We have been able to save about 10 to 15% on space using the space consumption features. 

The approximate cost per IOP is around 50 cents. 

I would rate this product a nine because of its ability to upgrade easily and the scalability.

Definitely look at the compression, deduplication, and the ability to grow into the cloud, consider any product that has these features. 

It's our main storage platform for our business applications. We have SQL and we do a lot of video editing. We also have a lot of media and sound data. Radio stations and TV stations keep some of their data on it.

Primarily, I do the data replication, the disaster recovery. Often someone will say, "Hey, I deleted this thing yesterday, can you get this back for me?" It's good that we have all those incremental Snapshots at different points in time that we can refer back to, to get them whatever they need.

Or we'll have someone download a virus, something malicious that corrupts a whole folder-tree of files. We can easily go back to right before they did that and just grab everything back. There are also ransomware attacks where they hold your data hostage by encrypting it. We can easily just go back and grab the data from before the attack. We can look back at the footprint and see when the whole tree was changed and restore the whole folder or the whole subset of folders. We might be down for two hours from it. An attack like that hasn't happened in the last nine months, but it has happened in the last two years.

It has also helped us reduce our overall cost of storage. Through dedupe and compression, we save a lot of capacity. Without that, we'd obviously have to buy more capacity. The last time I checked we were saving about 40 percent over our previous storage.

One of the things I find most valuable is the way they do the Snapshots, taking incrementals at points in time.

We just upgraded to 9.3, which is not the latest version, and it has some adaptive QoS built in. We have been using WFA for that. I'm interested in checking that out. I'm really glad they added it.

They use a lot of PowerShell for managing things and there are still a few things that you can't do through PowerShell cmdlets that you can do in the native CLI. It would be nice if they got more of those added in.

It's very stable.

It's very scalable as well.

The tech support is really good. We go through a support provider, Datalink, and those guys are really good. Anything that they don't have immediate knowledge of will be quickly escalated to NetApp and they're really quick about getting an engineer on it and getting us a solution. If we need a part replaced - of course, it depends what level of support we have for that particular system - everything is usually pretty quick.

In terms of a previous solution, when I joined this team they were already using NetApp. We did upgrade our controllers from 6290s to the 9000 but that was because of the age of the system. It was out-of-support and the support cost to maintain them got higher and higher as the years rolled by. It was cost-effective to invest in a new controller.

I wasn't part of this team when they initially set up all the storage, but we've had some upgrades, and we've gotten new controllers and added them to the cluster, and taken some out. All of those steps have been really straightforward.

Our reseller is Datalink and our experience with them has been really good. We've used them for several years. They negotiate really good prices for us and they give us really good support. If we need someone at a remote site, they'll schedule someone to support us at that site.

I don't handle the numbers, but the biggest ROI, to me, would be the ease with which we have our data protected.

We did look at Pure Storage and they were comparable in performance but they didn't have as many features. The gain in performance, for us, didn't offset the loss of the features, coupled with losing the relationship, or hurting the relationship, that we have with NetApp and Datalink.

The Pure system didn't have a way to do the iSCSIs that we need to use, and the impact to the relationship, to have a one-off system that didn't match the rest of them, didn't seem worth it to us.

In terms of NVMe over Fabrics, we no longer have any Fibre Channel. That was all phased out before I got on the team. In general, NVMe over Fabrics is good, it's quick. We aren't yet using machine-learning, AI, or real-time analytics but that is something we're looking into.

We have been using it for our entire environment, so we use it in file shares and in our VDI environment, at least for user data. That has reduced our user login times by about 50 percent. This has helped, and probably been one of the biggest benefits that we've recently seen, getting people's logins down to 45 seconds from more than two minutes. It's helped with our backups times, and keeping our backup windows in compliance.

The main thing has been performance, but we're hoping that we also get there on the reliability section (with the MetroCluster), so we don't have any disasters. The performance has been so good on it. Even some of their tests that we're doing, switching workloads from one data center to the other, has been great.

While its performance has been great, the configuration with the MetroCluster has been a bit of a headache. We've been working through it. However, it's a bit of a tedious process. With the MetroCluster, in theory, we should have good reliability. We have been having some issues just the finishing the configuration. Some of our DR tests have been not successful, so it's been a little troubling.

We have had some issues with the architecture, because of the complex setup with the MetroCluster. While it is getting there, we've had problems. We've had some outages trying to get the system up, so it hasn't been very good, but I'm hoping we get there. This does not seem like it's necessarily a NetApp issue, but the architecting of it has been a little rough.

Going forward, I want to see a simplified deployment and more straightforward recommendations on what is required for it. They're a little vague, and we've had some issues where we would like to have a tool that can scan the environment and find issues before deployment, fixing issues. So, when we're deploying it, we're not running into problems. Simplification would definitely help.

The system stability has been good. 

We haven't had it too long, so we haven't needed to scale it up. It does look pretty easy. We have MetroCluster, then we have the All Flash FAS and the new 8200 hybrid. 

We've been able to scale out the 8200, which has been beneficial, as we offload some of the older, slower storage to it. We have quite a bit of room to grow by adding new nodes into it. Right now, we only have two-node clusters, which have been great to upgrade, and not an issue.

Technical support has been good, but since we're using MetroCluster IP, we've been a little bleeding edge. This has been a bit of a problem because there are not as many people who are knowledgeable in it. We have had some issues getting people who have good knowledge on the subject, which has been why we've had some of the issues that we've had. A lot of it is just foundation and architecting. We've had some issues with existing equipment compatibility issues that we didn't really discover in the initial planning. The technical support has been good, but it has been a bit of a slow process.

We ran out of space with our old storage. We didn't have any full DR redundancy because we didn't have enough space to do a DR scenario. The main reasons to upgrade: Not enough space, we needed to upgrade, and NetApp happened to be a cheaper solution than some of the competitors.

We migrated from Dell EMC storage, where did have some issues with capacity in the IOPS, and NetApp was the cheapest solution. We got more performance and bang for our buck, so we moved over to the NetApp All Flash system. 

The initial setup was pretty simple, just for the basics. However, when we tried to scale out and run it in production, that's when we ran into issues. Because we were sold Fibre Channel MetroCluster, then that ended up not working for our situation, we had to buy more equipment and switch it over to MetroCluster IP. This was smooth at first, but then the deployment on production was much longer process.

All the support and deployment has been primarily from NetApp. We do have a consultant who has been assisting with some of the networking pieces, but most of the actual storage is all done with NetApp. While they've been knowledgeable, it's a complex configuration, and sometimes we've had some issues with them giving us solid answers on which direction we should go. They'll help us do something, but they don't want to give too much architecting information. That's been a bit of a headache for us. They said it was going to take 60 days, and it's been six months and we're still not fully rolled out.

We went through a reseller to buy the solution.

We looked at just Dell EMC and NetApp. Those were the two main choices, as they are similar products with similar performance. We could see Dell EMC was definitely more expensive, but it also had a more complex configuration. On paper, NetApp seemed cheaper and a bit simpler, which was one of the main reasons that we moved over. The performance for cost ratio was really interesting.

I would recommend doing your homework and make sure the environment is prepped before getting it set up. Make sure you have any prerequisites and your equipment's compatible: The distance between data centers is the networking pieces being compatible with the all-flash MetroCluster. If you have any issues there, you're going to have constant headaches with the configuration. The main piece is make sure you have all your ducks in a row on your networking gear, make sure it's compatible, and fix the issues early before you start deploying.