Storage Spaces Direct picks up the baton from Microsoft Clustered Storage Spaces, meaning it is a true SDS technology for Windows Server 2016. The testing of S2D is in full swing right now, and in this research, we provide a systematic guide on a practical deployment of a fault-tolerant 4-node setup.
Enterprises seek for ultimate performance, total fault-tolerance and availability to secure their business continuity. Thus, the 4-node setup which ensures exceptional uptime even in the case of multiple node failures seems to be the right choice.
The setup consists of 9 detailed steps: from Windows Server roles and features installation to the node failure test executed with our own specialized tool called Storage Data Corruption Test.
In the result, we have successfully built an S2D 4-node setup, which can withstand a node crash. It proves to be a good example of Enterprise-level infrastructure ensuring high uptime and performance due to overprovisioning in hardware resources.
Read the full article here: https://www.starwindsoftware.com/blog/microsoft-storage-spaces-direct-4-node-setup-2 to learn more about the setup process and detailed failover check results.
Storage Replica is a feature of Windows Server 2016, designed for the disaster recovery purposes. The replication it provides is block-level and volume-based. Storage Replica is targeted at disaster avoidance since the replication of data is done to a remote location.
Before, Windows had replication only at other levels, like file-to-file replication, applications replication, and VM-level replication. There was also block-level replication in the storage market, but vendor lock-in made it too expensive.
Storage Replica provides new opportunities for disaster recovery and disaster preparedness in the industry. With it, the data is synchronously protected at two separated sites, either different buildings, cities, or countries.
Storage Replica provides synchronous and asynchronous replication. With synchronous replication, the data is written to two locations simultaneously before completion of the IO, which is a way to go for mission-critical data. Asynchronous replication provides faster response time to application, so it comes in handy when the sites are too wide apart.
Read the full article here: https://www.starwindsoftware.com/blog/microsoft-storage-replica-in-a-nutshell
The article provides comparison of three leading products of the Software-Defined Storage market: Microsoft Storage Spaces Direct, VMware Virtual SAN and StarWind Virtual SAN. There are several use cases considered, based on the deployment scales and architectures.
Microsoft Storage Spaces Direct and VMware Virtual SAN are a perfect choice for bigger SMBs and entry-level enterprises because their licensing is reasonable for typical of these businesses infrastructure types. These solutions are not good with smaller SMBs and ROBOs, being too expensive and a performance overkill for them. The per-host licensing is too expensive for the hyperconverged environment of very big enterprises. Microsoft can expose only SMB3 reasonably well, while VMware “speaks” iSCSI and NFS, which prevents them from creating single shared storage pool in a multi-tenant environment. For the databases scenarios implementation, Microsoft and VMware have specific financial and technical issues.
StarWind Virtual SAN requires minimalistic two-node setup and provides 24/7 support, which makes it a perfect choice for small SMBs and ROBO. It has flexible licensing, to cover different deployment scenarios. StarWind utilizes the majority of industry-standard uplink protocols, so it can work with vSphere and Hyper-V environments simultaneously and provide a single pool of storage instead of separated “islands”. For datacenters, StarWind is good both in its software form as a “data mover” to create the virtual shared storage pool, and complete “ready nodes” for HCI or storage-only infrastructure. StarWind supports non-virtualized Windows Server environments, properly supports all possible storage protocols, and can provide high performance shared storage.
In general, StarWind Virtual SAN rather complements Microsoft Storage Spaces Direct and VMware Virtual SAN, than competes them. It fills the gaps for Microsoft or VMware-based infrastructures, providing them with the features to fine-tune different types of architectures.
Read the detailed comparison here: https://www.starwindsoftware.com/blog/software-defined-storage-starwind-virtual-san-vs-microsoft-storage-spaces-direct-vs-vmware-virtual-san
We’re making a series of tests, dedicated solely to the work of Resilient File System (ReFS – https://msdn.microsoft.com/en-us/library/windows/desktop/hh848060%28v=vs.85%29.aspx).
It is a relatively new proprietary file system, introduced by Microsoft in Windows Server 2012 as a successor of NTFS. Among its advantages over the predecessor, Microsoft lists enhanced protection from data corruption, common and silent alike, if provided with a redundant storage. The ReFS is also aimed at modern understanding of high capacity and large size. It supports files up to 16 million terabytes and maximum volume size (theoretical) of 1 trillion terabytes.
The tests are associated with workloads typical for virtualization – you can read more about them here. Our purpose is practical, because as of Windows 2012 time, Resilient File System has major issues with virtualization.
In the first part, we are planning to see how ReFS works with the FileIntegrity option on and off. This option is responsible for the data repair process, so it’s crucial that it works well under the pressure of random I/O, which dominates among virtualization workloads. You can check out the test here.
The second test is dedicated to performance because there have been reports of ReFS performance troubles somehow caused by hashsumming. We are about to see if the reports are true and if hashsumming has anything to do with any problems.
Development of any technology requires ideas, which need a lot of testing, before then can actually work. The problem is that testing and POC typically happen before the idea can make any money, relying heavily on free and open-source solutions.
There is no decent free SMB3 fileserver. Those available are critically unreliable or just aren’t working properly. In case you need one right now and there is no other way of acquiring it, there is a way – the Free Microsoft Hyper-V Server 2012R2. However this method violates license agreement, so it’s not in any way a permanent solution for your problems. You may take the risk and try it once in order to see if your idea works, but we still strongly discourage you from repeating this experiment.
In any case, let’s see if you can build an SMB3 fileserver on Free Microsoft Hyper-V Server 2012R2 and if it works, go further and see if we can create a failover fileserver. Check out the following posts and see:
Hyper-V: Free SMB3 File Server
Hyper-V: Free “Shared Nothing” SMB3 Failover File Server
Microsoft has always been targeting mainly SMB and ROBO space and now they’ve decided to aim at Enterprise, namely datacenters, Internet Service Providers, Cloud Hosting, etc. How well do they fare?
Problem was, up to Windows Server 2016, Microsoft had no real Software-Defined Storage, meaning it relied completely on SAS hardware. Not much “software defined” there. Enterprises couldn’t go with poor scalability of SAS with their several feet long cables and a literal headache of stretching the infrastructure as far as the next building.
A shared JBOD Scale-Out File Server (Image Credit: Microsoft)
In Windows Server 2016, Microsoft has designed a technology called Storage Spaces Direct, which is really SDS. What’s really good about it is the fact that the same engine can be utilized for various business sizes. There is no hardware lock-in, no distance or topology limitations. Having dropped hardware lock-in and accepting commodity components, the new technology also cuts down hardware expense and associated management costs as well. Besides, it is much easier for IT specialists to master one technology than dig into multiple different ones. How? Check out the full article here https://blog.starwindsoftware.com/microsoft-storage-spaces-direct/ and see how well S2D fares.
Let’s talk about the term “chain reaction”. Some of you may think only an unskilled administrator allows this to happen and that normal Fault-Tolerant setups cannot crash. Well, those of you are obviously wrong. Accidents happen and in worst-case scenarios, with multiple servers going into unplanned downtime, even an esteemed IT guru can find his system dead. Remember – there is always a chance higher than zero, and as Murphy’s law states, if something bad can happen, it surely will.
Microsoft came up with a nice idea, called “Soft Restart”, which is meant to shorten the time of the server’s vulnerable reboot state. It removes the hardware initialization phase, which couldn’t be shortened by any other means. Basically, Soft Restart only resets the software, so downtime is minimized, because software loading phase can be shortened dramatically with faster CPUs, flash utilization, etc.
We tested the alpha version of the feature and, sadly, it didn’t work in both physical and virtualized environments. Waiting for Microsoft guys to fix everything, because the idea may be a life-saver for someone. Check out how it all went for us here https://blog.starwindsoftware.com/2014/12/09/soft-restart-feature-the-great-idea-from-microsoft-its-a-minor-convenience-to-most-but-a-life-saver-to-some/