Pure Storage FlashArray Reviews

We use it for virtualization.

We have complete control over it. 

We partner with Oracle on stuff, so we have support on it. Before, we have had issues with support from our other vendors. It has been a major improvement from a support perspective.

The connections are a lot faster than what we had in the past. We've run Fibre Channel for our historical storage, and going from Fibre Channel to InfiniBand connected has blown up the amount of traffic that we can do. Now, we can support 40Gs per link, and there are eight links. When we went from being able to support four Fibre Channels at eight gigs a piece. One InfiniBand does what we did on all of our Fibre Channels.

The most valuable feature is the speed of it. It is much faster than anything that we can get from similar competitors.

The solution helps to simplify storage.

Historical analytics would be useful. At the moment, they don't have any type of application built for historical analytics.

The stability is very good. I've done destructive testing on it and never had any type of storage outages from it.

We haven't scaled yet, but we're planning to. We do upgrades constantly.

It's very scalable and easy to do.

We moved from an infrastructure that was owned by another team, so we needed something to move our own stuff onto. We originally tried some hyperconverged solutions from Dell EMC, but they didn't perform well at all. It took years to get that together and when we ran our benchmarks on them, and we decided they were not good. So, we immediately turned eyes to Oracle who is a big provider for my whole company, not just my team, and talked to them about what they had that was going to suit us, and they pointed us towards Pure Storage. As soon as we had a proof of concept and were testing it, we decided to run with it.

The initial setup was a little complex. We had some initial issues with the design and had to help correct some of the white papers for it, but it wasn't your standard use case.

We used Oracle.

I would recommend Pure Storage.

We investigated some flash storage implementations for it and based off of the way that the appliance works the added cost of flash doesn't scale with the performance that you get with it, so it hits on our middle ground. It works perfectly for us. We don't need to look at any type of flash storage.

We use Pure Storage for most of our databases, as well as for other application binaries.

Our biggest database is the Oracle ERP. Right now, it is around nine terabytes. It has grown from four terabytes in the last five to six years. Initially, we were worried because most of our functions and processes got slower. We thought we might have to add more  infrastructure and upscale it from a CPU perspective. Then, we moved to Pure Storage, and we suddenly saw some of the processes, which were running slowly, sped up automatically.

We have also seen a reduction in the latency.

The most interesting feature is the speed at which it executes I/O. After moving to Pure Storage, I have noticed that our databases are considerably faster.

Our performance has improved by at least four times.

The support for NFS protocols right out-of-the-box need improvement. I'm used to other storage vendors who have NFS support right out-of-the-box, and Pure Storage doesn't seem to have anything. We have shared APPL_TOP on our Oracle ERP, which would require an NFS type of storage. So, we had to resort to building our own NFS VM, then attach Pure Storage to it, and have it go through the server. This didn't really serve our purpose, as it's a lot slower because it's now going through a VM installer NFS server.

While we know Pure Storage supports snapshots, we haven't been able to implement databases or replication using them.

It has been very stable. I have not seen or heard of downtime storage issues after moving over to it.

It looks like it is scaling pretty well.

I have not used the tech support.

We have seen a reduction in total cost of ownership (TCO). 

Use Pure Storage for databases. I have seen a huge increase in speed and performance on our databases.

We use Pure Storage FlashArray as the storage for our virtual servers and our reporting databases.

FlashArray has allowed us to go all flash. There is no more worrying about how many IOPS our different storage tiers have and trying to buy more disks just to get higher performance. We simply don't worry about performance.

The best feature is consistently lower latency, even when IOPS crank up to over 75K. The product maintains submillisecond response time, which is incredible.

I would like to see a Nagios monitoring plugin which watches the health and performance of the system. The only one available just checks volume capacity. 

Our VDI with VMware Horizon include 100 VM for office and graphics desktop with nVidia GRID, the low latency and the high deduplication permit to reduce storage footprint, power consumption and to increase the consolidation on few VMware Host.

We have consolidated all the VSI (55 VM) and VDI (100 VM) in the same storage, improving performance and overall virtual desktop acceptance.

The most valuable features are extremely low latency, high IOPS with VMware, inline deduplication and compression.

We liked the non-disruptive downgrade from FA-420 (POC) to FA-405 in production and the non-disruptive upgrade from FA-405 to M20.

Reports of performance and LUN utilization could be improved. The VVol support is just released in GA.

Pure has global deduplication and the reports (Analysis tab - Capacity) are not clear, you don’t know how much Shared Space is used by each LUN but only the “Unique“ Space and the Snapshot Space.

I’d love to view the average, minimum and maximum performance in the reports (Analysis tab - Performance) but it is only graphics and you need to export data in CSV to find this information.

There were no stability issues. It is always on from the first LUN deployment to VSphere.

There were no scalability issues. NDU and NDD were very simple, without interruption or performance decrease for its cluster nature, active/passive.

Customer Service:

I would rate the customer service as very high.

Technical Support:

They are all skilled. They upgraded the storage firmware from remote every time we asked via a phone appointment.

We have used 3PAR for VDI and NetApp for VSI. The new solution has a very low price/performance ratio. With the price of the SSD upgrade of our NetApp FA3220 and the one year maintenance of the old 3PAR, we bought Pure Storage with a three-year support agreement.

The initial setup was very simple. In two hours, we started to use Storage vMotion.

The product was implemented directly by the vendor’s team and they are very skilled.

In two years, we will start to save. With their Evergreen Storage subscription, we can enjoy protection of our investment.

Everything is included, so there is very simple licensing. F

We evaluated EMC XtremIO and NetApp FAS.

Test it in your environment in a PoC or in a Try-and-Buy to check your deduplication and compression ratios.

Pure has global deduplication and the reports (Analysis tab - Capacity) are not clear, you don’t know how much Shared Space is used by each LUN but only the “Unique“ Space and the Snapshot Space.

I’d love to view the average, minimum and maximum performance in the reports (Analysis tab - Performance) but it is only graphics and you need to export data in CSV to find this information.

IPv6 is now fully supported, Windows Offloaded Data Transfer (ODX) is added in the latest firmware as supports for Docker APKG 1.0.1 and Dynamic Volume in the Purity Container Engine.

SAN solution for Tier 1 storage for VMware, Exchange, SQL, and physical servers; trying to fit as much production load as possible on them.

The speed to deploy or move VMs is ridiculous. By migrating our important systems to the Pure arrays, we have removed any storage issues and questions when there is a problem of "is it the storage?"

The deduplication and compression rates are beyond impressive. The amount of data that I have moved to it from legacy storage has enabled us to retire units that are three or four times the physical size.

• Ease of setup
• Migration to it
• Upgrading
• Support responsiveness
• Kick-butt mobile app
• On box analytics and vCenter plug-in
• The ease of creating a snapshot for testing or recovery.

I cannot stress enough the stability, speed, and awesome deduplication and compression rates. The amount of data we have moved to it has allowed us to remove units four times the size. The mobile app is kick-butt and support has been topnotch.

Mainly, just some nit picky stuff, like allowing servers and volumes to be grouped. Therefore, it would easier to work with them in the GUI.

None.

No issues.

It has been great; easy to work with to resolve cases or perform upgrades.

Yes, plainly put, it sucked and we wanted to get to a solution that was fast, easy to maintain, and reliable.

We had issues with it not handling the load and could never trust that it was serving data fast enough to not be causing issues.

Easy as pie.

In-house.

• Speed
• Time saved in management
• Availability
• No fingers pointing at storage as an issue.

It is a more expensive solution, but it is worth it. You are getting what you paid for. I never have to worry about its performance or if it is the root cause of an issue.

We looked at Nimble and EMC.

I never have to worry about its performance impacting the firm. It has made working with storage as easy and simple as it should be.

I had high expectations and they have met or exceeded each one of them. It was an exciting day when I finally got them up and running.

Production storage for a lot of virtual machines. As a service provider, it is very important to have something with this much performance.

• Cuts VM deployment down to seconds. 
• Cuts latency under MS. 
• Amazing performance. 
• Very stable; no worries about how much it can handle.

• Non-disruptive upgrades: You can upgrade at anytime without worry.
• The GUI is very easy to use and intuitive. 

I would like some form of QoS implemented. As a service provider, it would be beneficial to have it.

Our database administrators had to run some manual process twice a week since the disk performance of our previous storage unit was not able to respond to the requests fast enough.

This process took approximately four hours and it had to be done manually twice a week by the DBAs. After the implementation, the time to run the process was reduced to minutes and it did not require any manual intervention from our DBAs.

Performance, deduplication, compression, and fast response time for requests from servers and applications.

I have not been able to find one yet.

We did, but it was partially due to our environment. We were running outdated firmware in the HBAs for our HPE Blade Servers and an old version of vSphere that it is not supported by VMware.

We decided to keep these servers under with the previous storage array to avoid disconnects and system outages.

No, the unit that we bought contains 40TB of usable space and we are using 10TB so far.

Technical support is great. They will dive in deep with your team to figure out what is causing the problem for them and find the root cause.

I wish they could collaborate more with the other vendors internally, instead of us opening cases with Cisco, HPE, VMware, etc.

We have another platform using spinning disk 600GB 15K RPM SAS drives, but our applications and servers have grown so much that the storage was getting saturated with the requests from the applications.

We discussed the option to add more SAS disk and memory to the controllers of that array versus buying a Solid State Drive Array (SSDA).

We concluded that it would be more beneficial for our company to invest in an SSDA, and the results paid off.

The initial setup was straightforward and very easy to implement, as long as you have all the information that you need ahead of time.

For instance, IP addresses, iSCSI IPs and adapters, switch configurations and ports enabled, etc.

The price seems fair according to the market. We analyzed multiple All-Flash Arrays (AFA) in the market, but Pure came at the top in many areas.

The price was slightly higher than others, but competitive, if you consider all the other features that you get from it. I love the Evergreen model to replace any parts after three years with a newer part as part of your support contract. The licensing is based on your capacity.

We researched other products from Kaminario, NetApp SolidFire, Nimble Storage, EMC XtremeIO, and HPE.

I always recommend a company to start with a proof of concept. That way, you can test your applications directly with the unit. It is critical to get a baseline of the before with your current storage array and after with an SSDA or an AFA.

Redundancy and the fault tolerance of the platform are the most impressive. Also, this allows all the hardware and software updates to be completed while the system is completely online.

The internal garbage collection process has been fixed recently in some OS updates so it is more efficient but that could be just a little better. That’s it!

Everything related to the hardware or software update of the system(s) are active with absolutely NO downtime and NO service interruptions.

The m50 will be our production machine for a brand new property, we have almost finished building. I can personally state that the Pure Storage Flash Systems are the best built and Pure has the utmost professional customer care. Pure only focuses on Flash and they do it well.

The FA-450 pushes around 200,000 32k IOPS and with our load, which is not small by any means, we top-out around 50k-60k IOPS. It’s fantastic! We will soon upgrade both boxes to the newer m70 with some added capacity and all of which will take probably 30 minutes to complete.