This content is part of the Essential Guide: Working in a virtual machine environment: VM-aware storage
News Stay informed about the latest enterprise technology news and product updates.

Tintri VMstore upgrades quality of service, snapshot management

Tintri VMstore's operating system upgrade improves the vendor's quality of service to enable better performance of mission-critical apps.

Tintri VMstore operating system today was upgraded, improving its quality of service and management of snapshots and virtual machines (VMs).

The VM-aware storage vendor released Tintri OS 3.2, which gives customers the ability to designate minimum and maximum IOPS to specific VMs so performance levels can be balanced to favor mission-critical systems.

The new Tintri SynchVM allows users to navigate back and forth between snapshots without losing older snapshots or performance history. The Tintri Global Center 2.0 management portal offers the ability to do deeper health monitoring and apply group policies across 32 physical hosts and more than 100,000 VMs.

"There is a lot of meat to this launch, a lot of interesting and unique things to it especially with the quality of service (QoS) capabilities," said Scott Sinclair, storage analyst at Enterprise Strategy Group. "Tintri says you can put a minimum guarantee on this application so it always gets a certain level of IOPS. That's the idea because IT guys care more about having certain applications meet [service level agreements] SLAs."

Tintri's QoS levels can be rebalanced to favor mission-critical databases and applications, ensuring they always have their fair share of the needed resources and they are guaranteed a certain number of IOPS even while contending with noisy neighbor VM issues.

Previously, Tintri's QoS levels ensured all VMs had their "fair share of the performance and fair share of the pain" when resources were rebalanced among VMs, according to Tintri senior director of product marketing Chuck Dubuque.

"This release now gives you a VM policy that breaks that fair share," he said. "A customer can say 'I have a SQL Server and I don't want it starved for resources.' The rest of the VMs now get their share of whatever resources are left over."

The top-of-the-line Tintri VMstore T880 array can handle up to 3,500 VMs and about 140,000 IOPS, Dubuque said.

Dubuque described Tintri SynchVM as "snapshots on steroids." The software, which works at the VM level rather than the volume or LUN level, now offers the ability to freely move back and forth among snapshot versions without losing the most recent snapshot or the history tied to the snapshots. Most snapshot software comes with the tradeoff that you lose the most recent snapshots the father back you go among versions.

"Each VM in a Tintri system can have up to 120 snapshots," Dubuque said "You can go back two weeks and forward three days without losing snapshots, without re-provisioning VMs and losing the performance history of the VMs."

The latest Tintri Global Center now has the ability to go beyond monitoring VM stores for health checks. It can set management tasks like naming conventions of VMs and apply policies across server groups. It can also set snapshot and replication polices. Customers can manage dynamic collections of VMs based on group definitions and policies. Groups can span VMstores, hypervisor types and geographies.

"Those policies follow the systems even when they move from one VM to another," Dubuque said.

The Tintri OS 3.2 will be available in May to all Tintri VMstore customers. Tintri SynchVM will be sold as separately licensed software for Tintri VMstore. The Tintri Global Center 2.0 is free for Global Center customers and requires separate license for other Tintri VMstore owners.

Next Steps

Why not to use LUNs for snapshotting

Tintri adds global management to its VMstore appliances

Dig Deeper on Storage vendors

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Is it important for your storage array to be VM-aware?
QoS is very important. Need to be able to define both floors and ceilings to be able to relate to SLAs for customers.
Having storage VM-aware saves a ton of time in the troubleshooting and monitoring aspects.  
VM-aware storage isn't just dumb storage. It actually improves performance of your VDI implementation.
QoS is an improtant part of both test and production environments. It allows for planning and scaling dynamically. 

It also allows for an equal resource provisioning system in environments where hungry boxes dominate the storage calls
VM level storage management has been a defining moment in computing for me.  It gives an incredible yet necessary level of control and visibility on what is happening on the ground floor of my virtual machines.   Taking vm management to the next level with QOS is an interesting twist and demonstrates that Tintri is a thought leader in this space, where other storage vendors are still playing in luns and volumes.
Yes, managing volumes and LUNS just doesn't cut it anymore.  We need a more granular construct to better manage our virtual machines...the virtual machine!  Tintri is leading the way in this respect.  
Coming from an MSP shop using "traditional SANs" I was blown away by the performance increase vmaware storage provides. 
Troubleshooting and finding the offending machine is a simple affair now.
This is a great example of the benefits of Vvols; with something like this you can enforce a QoS for a mission critical app and possibly back it up during operation hours with no impact.
Absolutely. This is one of the main reasons I went with Tintri. Using NetApp as a SAN when trying to manage VDI, 2+ TB Oracle DBs, SQL DBs, Exchange, etc. is just impossible. All the LUNs impact the overall performance of the aggregate and I can never get the IOPS/Latency requirements I need. Not to mention once the aggregate starts creeping above 80% capacity, I start seeing a drop in performance and additional latency. With Tintri's VM-aware, I just put everything into a single NFS datastore and let the Tintri do the performance tuning. And since Tintri also allows QoS and pinning individual VMDK files, I can fine-tune the performance even more and make sure my developers are not hogging the IOPS. :-) Did I mention how valuable the performance charts are? Now I can easily see what VM has the highest latency and where that latency is coming from, and 9 times out of 10 it is coming from the ESXi host. 
"The Tintri OS 3.2 will be available in May to all Tintri VMstore customers."

So did this ever ship? 
You can get the latest OS by logging into
I'm interested in talking to some users/customers who posted to this article's discussion section. I'm working on a story on this topic. I can be reached at 339-222-3103