Accelerate DevOps with Copy Data Management and Tintri SyncVM
Accelerate DevOps with Copy Data Management and Tintri SyncVM
Learn how a Tintri customer used copy data management with Tintri SyncVM to reduce SLAs and facilitate a transition to DevOps.
Traditional storage impedes the provisioning and restore processes needed for effective DevOps.
A financial services customer reduced SLAs for common tasks from five hours to five minutes with Tintri.
Tintri SyncVM provides copy data management, facilitating provisioning, providing fast and efficient restores, and enabling time travel.
The financial services industry is rapidly adapting to the digital economy. Bank customers rely on web services and mobile applications to access data and make transactions and new customer workflows continually test the limits of existing code.
At Tintri, we’ve been hearing a lot from financial services and other customers that want to accelerate development and are aggressively adopting a DevOps model. Given the rapid changes, many are struggling to keep up with developer needs using traditional storage processes. Tintri can streamline your transition to DevOps by shortening and simplifying time-critical tasks across your Prod, Dev, Test, and QA environments. Tintri copy data management eliminates physical data copying and lets you use storage more efficiently.
Operations to Developer Needs
One financial services customer runs a vSphere 6.0 and vCloud Suite environment with over 700 production VMs running customer-facing applications and 800 development VMs.
Given the increasing reliance of the company on web and mobile, an hour can seem like a lifetime to a developer waiting for resources to troubleshoot a critical bug. The Dev team came to the Ops team with three big asks:
Standard SLAs for provisioning. Developers wanted guaranteed provisioning times for VMs with the latest production data.
Standard SLAs for restores. Developers needed to know how long it would take to restore a VM, a vDisk, or a file.
Guaranteed response time on service requests. Even if a task can be accomplished in minutes, it doesn’t help the developer if his or her request sits in a queue for two days. Response times to requests varied widely depending on staff availability and the current system load.
With Tintri, the Ops team was able to address these requirements and decrease the SLA for provisioning and restores from five hours to five minutes. Tintri storage gives the Ops team three key capabilities that save both time and resources:
VM and vDisk refresh with up-to-date data from production within minutes. This ensures that Dev and Test are never out of sync with the production environment.
Restores down to the file level without consuming additional storage. Previously, restores could be delayed due to limited space, necessitating a frantic cleanup. That’s no longer an issue.
Time travel between different point-in-time versions of a VM. To understand and troubleshoot problems, Dev sometimes needs to go back to earlier versions of VMs, complete with the data set as it was at that time. Tintri not only makes that process possible, it makes it simple.
These capabilities allow the Ops team to be much more efficient. Because they can automate many tasks—and accomplish in a few mouse clicks what used to take hours—request queues are much less likely to back up.
Fast VM Refresh with the Latest Production Data
In development environments, it’s important to refresh with the latest data from production as often as possible. Ensuring that Dev is a close match to Prod is one of the best ways to avoid bugs and integration problems.
The problem with most approaches is that you have to rebuild the entire environment, including all the necessary VM settings, any time you refresh the data. This makes the process cumbersome, time consuming, and error prone.
With SyncVM copy data management, Tintri makes it possible to accomplish refreshes without rebuilding the entire environment, streamlining the process. As a result, there’s less disruption and you can refresh more regularly.
SyncVM solves these issues by leveraging snapshots and cloning instead of physical data copies. Because flexibility is key, SyncVM can be invoked at either the VM or vDisk level.
Our financial services customer was able to automate the process using Tintri APIs so that Dev can initiate the process themselves if needed without having to get Ops involved.
In a DevOps environment, there are any number of reasons you might need to restore a VM, a vDisk, or an individual file. Unfortunately, with traditional storage, that turns out to be a challenge. If your backups are at a LUN-level, that usually means restoring an entire LUN, even if you just need a single file. And finding enough spare storage space to restore an entire LUN is often a challenge.
With Tintri VM-aware storage, the process is greatly simplified. You never have to worry about operations at the LUN level; everything happens at VM or vDisk granularity.
Tintri space-efficient snapshots serve as a first line of defense. VM and vDisk restores can be handled manually using SyncVM. You can easily restore an entire VM or a vDisk or set of vDisks for a particular VM to the time of any previous saved snapshot for that VM. To restore a single file, you just attach the snapshot to an existing VM as a temporary disk and copy the file or files needed.
For routine operations, like rolling back to baseline after a test complete, these types of restores are easily automated using Tintri APIs. For unexpected occurrences, the Tintri console allows you to accomplish what you need with a few clicks.
As you accelerate the delivery of new code and move towards continuous delivery (CD), you’ll probably end up needing to move your Dev environment backward in time more and more frequently. For example, suppose Dev gets a report of a customer problem that occurred two days ago. To reproduce the problem you may have to restore some part of your development environment—including the production data set—to that point in time. Tintri makes that kind of time travel simple.
With Tintri, time travel is really just an extension of our efficient restore capabilities. Using SyncVM, you can restore a single VM or a set of VMs to the time of any saved snapshot. You can save hundreds of snapshots per VM, move forward and backward in time as needed, and have multiple developers who are working on the same problem all looking at different points in time.
Other vendors’ snapshot capabilities have significant limitations. Most critically—with some implementations when you make a previous point-in-time snapshot active—you sacrifice all the snapshots that come after it. You can’t do the kind of time travel that Tintri enables without jumping through a lot of hoops.
Smarter Storage for DevOps
For this financial customer, these capabilities have made a huge difference. SyncVM has enabled them to go from five-hour SLAs to less than five minutes. The Ops team can recover any VM, vDisk, or file to any point in time in a matter of five mouse clicks. Tintri has enabled the team to get its weekends back and they are now expanding to automate more processes and implement the same model in more data centers.
Dhiraj is a Director of Product Marketing at Tintri, responsible for enterprise cloud-specific products and solutions. In this capacity, Dhiraj speaks about using infrastructure to enable agile dev...more