We often see IT infrastructure vendors use terms like “Simple” and “Easy” to describe key traits of their next-gen products. Creative marketing minds work tirelessly to come up with catchy messages like “Radical simplicity” from Nutanix and “Simplicity at every stage” from Pure Storage and compare their user experience to that of complex legacy solutions to prove their claims.
This all sounds great on the vendor websites, but simplicity and ease-of-use are difficult to measure. As we know, IT operations reality is usually a far cry from high-level marketing claims – which brings us to the following true story recently told by a joint Tintri and Pure Storage customer:
Out of the blue, the customer noticed that more and more space was being consumed on their Pure array. The array was filling up faster than usual and it seemed as if reaching 100% full was imminent – threatening outages of production workloads. There was no time for Storage vMotion – Pretty scary!
According to the customer, they turned to the Pure Storage array’s user interface, but it couldn’t help them identify the guilty app, VM or even the rapidly-growing volume. Colorful array-level graphs are not useful in identifying trends of specific VMs when trying to troubleshoot an issue and isolate a root cause.
With the help of the vendor, they resorted to using vendor provided scripts and examining log files to finally find the guilty volume. Having the volume information allowed them to check vSphere and find the root cause – a home-grown application VM writing huge amounts of data to disk.
This story is a great example that no matter how “simple” and “easy” a product claims to be, if the product has known technology limitations, it cannot deliver on its promise. The Pure array is simply missing VM or application level information.
Since this customer also owns several Tintri All-Flash arrays, they know that in a similar situation with Tintri as the storage, a single glance at their Tintri Global Center would have given them the answer.
As seen in the following screenshot, Tintri shows the top 10 space changer VMs along with the top 10 performance reserve changer VMs, right there on the main page:
This capability is so powerful and yet so seamless to use, that many of our customers (and even us here at Tintri) often take it for granted – until we’re reminded how difficult it could be to manage a production environment while having no visibility. Running blind, hoping that VMs will play well together in their shared volumes, is not a strategy.
While traditional approach is familiar and comforting at times, this simple example shows why traditional approach to storage is not suitable for hosting virtual machines.
The ideal solution is one that leverages deep integration with the hypervisor layer and a storage array level file-system that was designed from the ground up to manage objects such as VMs and containers, rather than LUNs and volumes. If you’re tired of running blind, take a look at how Tintri can help.