No matter the brand, type, size or performance of you storage system, they should all have one thing in common: stability. Pure Storage talked about their systems design at Storage Field Day 8, which centers around “non disruptive everything”. Not only the hardware, but also the software running on top of it. Because in an ideal world, the storage system should only go down 5 years after your installation date when you’re replacing it with a new one. And in Pure’s case this means: never.
Monitoring with Pure1
It all starts with monitoring. Proactively solving tiny issues before they could become big ones that cause downtime or even worse, data loss. Pure Storage does this by automatically collecting telemetry from all their FlashArray//m systems in the field and aggregating all that data into their Pure1 management and support cloud.
This enables two things: first of all you, the customer, can monitor and manage the storage system from all over the world, be it with a browser, a tablet or your phone. Additionally, with the thousands of systems in the field all sending data back to Pure, they can see trends on these systems before a single company or engineer would realize their system is at risk. The Pure support department automatically generates support tickets and correlates and fingerprints them. This helps to prioritize and solve potential system problems before they can cause any disruption for the customer.
Pure will also handle your system upgrades. Of course the customer is consulted beforehand: only with customer permission will Pure Storage upgrade your array remotely. Most of these upgrades should be able to be installed without performance impact, i.e. you could upgrade during business hours. This works so well that of the 1000 arrays shipped last year, 91% is already running an upgraded software set.
Apart from hunting for potential issues the Pure1 GUI can of course also be used for usage reporting and forecasting; with forecasting being configurable to meet your own forecasting period: for example for your yearly tech refresh/capacity expansion.
The Pure Operating Environment (Purity) runs on fully modular & redundant hardware. During the presentation at Storage Field Day 8, Pure demonstrated this by pulling out various components out of a running system: power supplies, controllers, flash modules, everything. All non disruptive for the simulation workload on it, and also without major performance consequences. There’s a video on the Tech Field Day website, just follow the link further down below in this post.
In case of a drive/flash module failure, a rebuild will automatically start. There are no hot spares in the system; instead the rebuilt data will consume free space from the system. If you’ve pulled a drive from the system and reinserted it later on, the data on that drive is automatically re-ingested. The metadata in the system will make sure that you’re always using the latest copy of data: old, outdated data that’s re-ingested will automatically be discarded and returned as free space.
So earlier in this post I’ve said that Pure systems shouldn’t necessarily have to go down after 5 years to make room for a new, refreshed system. This is because the FlashArray//m systems can do a non disruptive chassis upgrade: one controller is pulled out, the remaining old controller is connected to the new chassis, data is moved and voila! Pure calls this Evergreen Storage: no more forklift upgrades where you have to migrate the entire environment when you replace the hardware. Pure even promises a 10+ year lifespan!
My thoughts on Pure Storage
First of all I’m glad to see that Pure is still heavily investing in monitoring and support. My day job still has plenty of post-sales work, which means I do my fair share of software upgrades, installs and configurations. Installing systems made by a company with a solid focus on support is golden: yeah, stuff breaks… and in that case you need a quick and professional response. Several of the SFD8 delegates were pretty positive about their own experiences with Pure Storage, which is definitely a good sign.
The FlashArray//m systems are scale-up systems. Pure did this on purpose: their customers want multi-TB systems, not multi-PB systems. Pure has explored scale-out during the initial design phase of their products, but rejected it due to the added complexity to the product. I’m curious to see if Pure will jump on the Silicon Valley scale-out bandwagon and add scale out functionality in the future: fellow SFD delegate Chris Evans clearly seems to think so, as you can see in his post about the FlashArray//m platform.
Are you a Pure customer? Share your support experience and what you think of the Pure1 GUI!
Disclaimer: GestaltIT paid for the flight, hotel and various other expenses to make it possible for me to attend SFD8. I was however not compensated for my time and there is no requirement to blog or tweet about any of the presentations. Everything I post is of my own accord.