Our biggest bump was going from the 4x software to 5x software and getting them to understand the nuances of how we did real-time replication. They have an automated failover that requires a third-party server to be a watchman over the storage cluster, which is very common. When we were doing that upgrade, it turns out we didn't need that. We were doing manual failovers but we still installed that watchman because we thought we needed it to do our upgrade from 4x to 5x. But once we got it up to 5x, we never had to tell our existing replication settings to use the watching node and to do automated failover. It took multiple phone calls and conversations until we finally got to our local sales engineer who has known the product for a longer period of time than some of the support people we were dealing with to clear up that inconsistency.

Tintri's Cloud Connector currently only goes to AWS and IBM Cloud, and we don't use either because we're Microsoft Silver Partners. It would be great to get the Cloud Connector feature with Azure. If it's not already on Tintri's roadmap, that's something I'd like to see.

While the performance and capacity monitoring GUI is excellent, I'd like to see multiple windows simultaneously. You can see various statistics, but it would be great to have that across several frames to monitor more things simultaneously. That's a minor issue. 

The biggest area for improvement, and there has been some roadmap work in this area already, is cloud integration. It would be hard to find an organization that isn't actively embracing the cloud, at least in a hybrid way. Some workloads are very well suited for public compute on Azure, AWS, GCP, et cetera. That's important for us and for our customers, but it's also important from an OpEx perspective. 

Tintri has been investing in this area and I'm sure will continue to, but cloud integration has been the biggest area that we've been crossing our fingers and hoping for quick development around. 

We're a Microsoft Gold partner and a lot of our customers have heavily embraced the Microsoft ecosystem. To do modern work securely, public compute and productivity on the Microsoft Office 365 stack, Exchange Online, SharePoint, Teams, OneDrive, et cetera, are important for mid-market and smaller organizations. There has to be a deep integration point between the on-premises solution, a hybrid cloud solution, and a public cloud consumption model. That is becoming an increasingly important decision point in all of our vendor-alignment discussions, Tintri included.

I'm waiting to see the Kubernetes package. I know they're releasing one, but I haven't seen it yet. So, I'm waiting to see that. 

As long as they stay, I don't want them to rest on their laurels. They're been great. I want them to continue innovating. The way it is VM-aware storage, it'd be nice to see cluster-aware storage as well. That'd be cool, and I know they're working on that.

Tintri has a feature called Tintri Global Center, which helps to monitor and control our infrastructure with a single click. This works fine if you want to look at existing data, however, if you want to monitor dynamically with a granularity of seconds that it requires a connection back to Tintri central, which is an extra step and not so convenient to use. Also, we have noticed that while Tintri has improved our infrastructure performance by 30%, the performance drops significantly if we exceed SSD capacity.

Following additional features, we would like to see in the new release:

1. The fixing of the performance issue when we exceed SSD capacity.

2. Documentation support in major languages besides English (as we have teams spread out across multiple countries).

3. Having a one-click GUI to create snapshots of the storage.

4. Detailed reporting is missing in the current version. We would like to see this feature added in a new release.

It would be beneficial if Tintri could identify issues through telemetry and notify us before we need to contact them.

The service that Tintri provided based on the telemetry data was quite useful. However, the only time maintenance would occur was when we had to manually investigate the issue, whether it was related to replication, controller failure, or failover. This is the one area where I believe there could be improvement.

We need more options to integrate with cloud storage options other than the current AWS and IBM that it currently supports. We also use Microsoft Azure services for a portion of our database and would like to see the seamless integration of that platform with Tintri VMstore as well. 

Additionally, support ticket times could be reduced, and the method could be made a little more robust to save time and effort. 

The addition and improvement of these features can make Tintri VMstore stand out as the leading solution for VM storage and Database management.

The Tintri OS and GlobalCenter software do a great job of showing you troubled VMs, however it still could be a bit more helpful in diagnosing the issues. We have had a couple Linux VMs that went out of control with IOPs and Tintri showed us that, but we still had some digging to find out why. So adding some more detailed information as far as where the IOPs issue is coming from would really help.

I have not had any major problems with it except that one of the controllers failed one time. But that didn't hinder performance because there were two of them.

A feature like replication for disaster recovery would be nice.

I would like it to have the ability to store data other than virtual machines. At the moment, you can only connect VMs to it, and that’s a bit disappointing.

There are occasional bugs but Tintri has been good about fixing them quickly. The bugs are mostly cosmetic; I've never had a bug impact end users or VMs.

In regards to the bugs, we have had minor issues here and there, when accessing things in the UI. Nothing to note recently but in previous releases, we have had a lot of issues with AD integration. However, those issues have been resolved.

We have also had false-positive alerts, while upgrading the code. That only happened with one version of code, that we upgraded to but again, those have been resolved.

Tintri is always very good with resolving bugs, i.e., by frequently providing code upgrades.

The area it needs to improve is the pricing. It could be better. Another topic is the replication feature, In the version 4.3 the syncronous replication is not so friendly and you have to do procedures manually to come back once you recover one of your nodes. 

I would love an Active/Active set-up, if possible.

Tintri hardware is perfectly sized for the medium-to-large office. Their interface and methodology is great, but it's a shame there is nothing for a robo office with 10-20 machines.

• I would love more insight into each virtual machine statistic. 
• The Tintri Analytics site is excellent for long-term trending, but more data would be great.

On the physical side, I would suggest adding a light to each of the controllers that would indicate which is the active and with is the passive.

On the software side, I suggest adding integration with backups software like CommVault Simpana or Veeam where these products can integrate with the array and trigger things like snapshots for backup, etc. We are considering buying another to host more of our virtual server workload.

Replication between two VMs needs to be improved to real-time. The solution also needs to have a better licensing model.

I would love to see improvements in their replication engine. Their current replication is really just enough to "check the box" that they do replication. We'll probably implement Actifio, Zerto or EMC RecoverPoint for VMs for more critical data replication.

Reduce the weight of the box so it's not such a bear to install in the rack. (More of a joke because I can't think of any improvements that can be made for performance and to the management interface.)

I would like to see better integration of the Tintri features directly into vSphere. Some of this is an extensibility issue in vSphere. They need to keep improving single pane of glass management across many Tintri units and not just monitoring of all units. They are making progress on both of these issues.

• Active/active cluster between two Tintris on Hyper-V cluster.
• Dedup should be better.

The product could be improved by adding iSCSI support, as others have mentioned. We have had to rethink how we implement some of our services due to this.

I think with the world soon becoming only SSD, possibly NVMe, and 3D Xpoint. It would make sense for Tintri to drop the hybrid array down the line. The storage market is becoming cluttered.

Backup integration with Veeam. I would really like to see a way to leverage storage side snapshots for this.

It needs current RHEV-M support (v2.5) which is slated for 2016. It also needs dedupe on the spinning media for the hybrid array. Change the UI to use something other than Adobe Flash.

None come to mind; the unit was a breeze to setup and I haven't come across any sort of setup issues at all.

Of course, a lower price would always be welcome. We were one of the first hyper-v adopters, so it was a bit more cumbersome for the initial setup, but the new OS has resolved these issues.

In sync and automated mirror between two Tintris is missing. However, as far as I know, it is on their roadmap.

iSCSI could be a nice add-on. Dedupe is a nice to have! At this moment, compression does a very good job, but I'm sure dedupe could be the next big thing for VMStore.

Speed of our VDI machines. We have a very high log in and log out ratio and machines are being refreshed instantly so we have a constant boot storm on our storage. Tintri can easily put up the effort then run this smoothly.

Integration of all the interfaces:

• For each Tintri
• For the global centre management
• Analytics

The interfaces are all great, but if all the functionality were built into one global centre, this would make the interface a single pane.

We have tested the replication piece – it actually works well, but is expensive based on our limited needs – if we were utilizing the Tintri for more of our VM workload it may be more cost effective.

It lacks deduplication, and dedupe on all data is something that would be a big improvement that we need.

Always could be cheaper! : )

I think that Tintri's cautious approach to introducing new features has been great however it has allowed competition to catch up to it. 

Tintri need to be able to innovate faster but maintain the quality of their features.

They should keep up the great innovation and good work.

In the end, we decided the price wasn't going to work for us, along with the fact there was fussiness around the storage. 

Technical support needs improvement as well.  

It needs a Hyper-V toolset and OpenStack Horizon support. Most of the features are in line with the competition, but they could provide automation to move VMs between hypervisors.

That’s tough. The thing is rock solid and easy to manage. Can’t think of anything right now, but reserve the right to revisit this.

The licensing model, however, is disappointing as each feature requires a license.

Stretched Cluster also for older models like T540 and T650.

Bye bye managing storage array, to create volumes and taking care to isolate big IO consumers into the production environment. Huge improvement in creating adhoc labs, pre-prod environments and dev as well, in just few minutes, etc.

Of course price can always be lower. Still working with support to enable more options for backup at this time, but we asked for features and they expect to deliver in the next month or two...nice to have a company that listens.

I would like to be able to add more storage capacity to our 2 units down the road with out buying an additional seprate array. I hear that their newer models may solve that issue though.

Right now, we see that the configuration access isn't as good as it should be. An improvement in this area would be a huge help for us.

More cloud integration.

It would be very helpful if it had scheduled time-to-full trend reporting because management always wants to know how much time is left on storage.

Integration with Commvault for native storage snapshots

Not a fan of SyncVM being an addition license. Dedupe is only at the clone level; no overall dedupe process. 

Make sure to do a POC. Then after that, put into production. You'll see that the technology is good. 

I would rate it an eight out of ten. 

The features are good but the brand name is not so popular here.

The Hyper-V configuration is not as easy as VMware. Documentation could be better as well. The replication technology is good, but could be improved to allow the powering on of the replica instead of having to clone the machine.

Synchronous Replication