Continuous available general purpose file shares & ReFSv3 provide high available backup targets

Introduction

In our previous two blog posts on Veeam and SMB 3 we’ve seen how and when Veeam Backup & Replication can leverage SMB Multichannel and SMB Direct. See Veeam Backup & Replication leverages SMB Multichannel and Veeam Backup & Replication Preferred Subnet & SMB Multichannel.The benefits of this are more bandwidth, high availability, better throughput and with RDMA low latency and CPU offload. What’s not to like, right? In a world where the compute and networks need keeps rising due to the storage capabilities (flash storage) pushing the limits this is all very welcome.

We have also seen earlier that Veeam B & R 9.5 leverages ReFSv3 in Windows Server 2016. This provides clear and present benefits in regards to space efficiencies and speed with many backup file related operations. Read Veeam Leads the way by leveraging ReFSv3 capabilities

When it comes to ReFSv3 in Windows Server 2016 most of the focus has gone to solutions based around Storage Spaces Direct (S2D). That’s a great solution and it is the poster child use case of these technologies.

But what other options do you have out there to build efficient and effective high available backup targets creatively except for S2D? What if you would like to repurpose existing hardware to build those? Let’s take a look together at how continuous available general purpose file shares & ReFSv3v3 provide high available backup targets

CSV, S2D, ReFSv3 & Archival Data

In Windows Server 2016, traditional shared storage (iSCSI, FC, Shared SAS, Shared RAID) with CSV are not recommend to be used with ReFSv3. Why isn’t exactly clear. The biggest impact you’ll see is the performance difference when not writing to the owner node of the CSV in this use case. Even with a well configured RDMA network that difference is significant. But that doesn’t mean that the performance is bad. It’s just that many of the super-fast meta data operations are relatively and significantly slower when compared each other, not that any of these two are slow.

clip_image002

Microsoft does state that an S2D with ReFSv3 and SOFS shares can be used for archival data. Storage spaces and ReFSv3 also have the benefit of offering automatic repair of corrupt data from a redundant copy on the fly even when needed. So yes, the best know supported scenario is this one.

Continuous available general purpose file shares and ReFSv3 provide high available backup targets

But what if we need a high available backup target and would love to leverage ReFSv3 with Veeam Backup & Replication 9.5? Well, you can have 95% of your cookie and eat it to. All this without ignoring the cautions offered.

We could set up SOFS shares on a Windows Server 2016 Cluster with ReFSv3 with traditional shared storage. Some storage vendors do state this is supported actually.

That only means you don’t have the auto repair functionality ReFSv3 combined with storage spaces offers. But perhaps you want to avoid the risk of using ReFSv3 with CSV in a non S2D scenario all together. What you could do is forgo ReFSv3 and use NTFS. How well this will work for archival data or backup is something you’ll need to test and find out how well this holds up. There is not much info is out there, only other cautions and warnings that might keep you up at night.

There is another scenario however and that is using Windows Server 2016 failover clustering to set up continuously available general purpose file shares that leverage SMB3 transparent failover.

The good news is that general purpose file shares (no CSV) do work consistently with ReFSv3 because such a share/LUN is only exposed on one cluster node at the time, the owner. By having multiple shares and setting preferred owners we can load balance the workload across all cluster nodes.

Thank to continuous availability for general file shares and SMB 3 transparent failover we can still get a high available backup target this way. The failover is fast enough to make this happen and all we see with Veeam Backup & Replication is a short pause in throughput before it resumes after failover. To put the icing on the cake, you can leverage SMB multichannel SMB Direct for both backup and restores.

I would take a sizeable whitepaper to walk through the setup so instead I’ll show you a a quick video of a POC we did in the lab here https://vimeo.com/212886392.

clip_image004

If you want to learn more come to the community & other conferences I’m speaking at and will be around for Ask The Experts time opportunities. I’ll be at the German Hyper-V community meet up, The Cloud & Datacenter Conference in Germany 2017, Dell EMC World 2017 and last but not least VeeamON 2017 (see  May 2017 will be a travelling month). 

Conclusion

What do you lose?

Potentially there is one big loss in regards to the capabilities of ReFSv3 with this solution when you are not using storage spaces. This is that you lose the capability to automatic repair of corrupt data. The ability of ReFSv3 to do so is tied into the redundant copies of Storage Spaces (parity/mirror).

What do you get?

That’s fine, the strength of this design is that you get the speed and space efficiencies of ReFSv3and high available backup targets in way more scenarios than “just” S2D. After all, not everyone is in a position to choose their storage fabric for backup targets green field or at will. But they might be able to leverage existing storage and opt to use SMB 3 for their data transport.

So even if you can’t have it all, you can still build very good solutions. It offers ReFSv3 benefits and high availability for your backup target via transparent failover with SMB transparent failover on continuous available general purpose file shares. This also only requires Windows Server 2016 Standard Edition, which is a cost saving. You get to leverage SMB Multichannel and SMB Direct. All this while not ignoring the cautions of using ReFSv3 in certain scenarios.

On top of that, if you use NTFS with this approach it will also work for Windows Server 2012 (R2) as the OS for the backup target cluster hosts.

Disclaimer

I do not work for or at Microsoft, nor am I perfect or infallible just because I’m an MVP. You’ll have to do your own testing and validation. From our testing and without ReFSv3 bugs ruining the show, to me this is a very valid and cost effective approach.

Veeam Backup & Replication Preferred Subnet & SMB Multichannel

Introduction

In a previous blog post Veeam Backup & Replication leverages SMB Multichannel post we showed that Veeam backup & Replication leverages SMB multichannel when possible.

But what about Veeam Backup & Replication Preferred Subnet & SMB Multichannel, does that work? We mentioned that we wanted to answer the question what happens if we configure a preferred back-up network in Veeam Backup & Replication. Would this affect the operation of SMB multichannel at all? By that I means, would enabling a preferred network in Veeam prevent multichannel from using more than one NIC?

In this blog post we dive in to that question and some scenarios. We actually need to be able to deal with multiple scenarios. When you have equally capable NICs that are on different subnets you might want to make sure it uses only one. Likewise, you want both to be used whether they are or are not on the same subnet even if you set a preferred subnet in Veeam. The good news is that the nature of SMB Multichannel and how Veaam preferred networks work do allow for flexibility to achieve this. But it might not work like you would expect, unless you understand SMB Multichannel.

Veeam Backup & Replication Preferred Subnet & SMB Multichannel

For this blog post we adapt our lab networking a bit so that our non-management 10Gbps rNICs are on different subnets. We have subnet 10.10.110.0/24 for one set of NICs and 10.10.120.0/24 for the second set of NICs. This is shown in the figure below.

image

These networks can live in a separate VLAN or not, that doesn’t really matter. It does matter if to have a tagged VLAN or VLANS if you want to use RDMA because you need it to have the priority set.

We now need to configure our preferred network in Veeam Backup & Replication. We go to the main menu and select Network Traffic Rules

clip_image004

In the Global Network Traffic Rules window, click Networks.

clip_image006

In the Preferred Networks window, select the Prefer the following networks for backup and replication traffic check box.

clip_image008

Click Add. We use the CIDR notation to fill out our preferred network or you can use the network mask and click OK.

To prove a point in regards to how Multichannel works isn’t affected by what you fill out here we add only one of our two subnets here. SMB will see where it can leverage SMB multichannel and it will kick in. Veeam isn’t blocking any of its logic.

So now we kick of a backup of our Hyper-V host to our SMB hare target backup repository. We can see multichannel work just fine.

clip_image010

Below is a screenshot on the backup target of the backup running over SMB multichannel, leveraging 2 subnets, while having set only one of those as the preferred network in Veeam Backup & Replication

clip_image012

Look at my backup fly … and this is only one host that’s being backup (4 VMs actually). Have I told you how much I love flash storage? And why I’m so interested in getting ReFS hybrid volumes with SSD/SATA disks to work as backup target? I bet you do!

Looking good and it’s easy, right? Well not so fast!

Veeam does not control SMB Multichannel

Before you think you’re golden here and in control via Veeam lets do another demo. In the preferred network, we enter a subnet available to both the source and the target server but that is an LBFO (teamed) NIC with to 1Gbps members (RSS is enabled).

clip_image014

No let’s see what happens when we kick of a backup.

clip_image016

Well SMB multichannel just goes through its rules and decided to take the two best, equally capable NICs. These are still our two 10Gbps rNICs. Whatever you put in the preferred network is ignored.

This is neither good or bad but you need to be aware of this in order to arrange for backups to leverage the network path(s) you had in mind. This is to avoid surprises. The way to do that the same as you plan and design for all SMB multichannel traffic.

As stated in the previous blog post you can control what NICs SMB multichannel will use by designing around the NIC capabilities or if needed disabling or enabling some of these or by disabling SMB multichannel on a NIC. This isn’t always possible or can lead to issues for other workloads so the easiest way to go is using SMB Multichannel Constraints. Do note however that you need to take into consideration what other workloads on your server leverage SMB Multichannel when you go that route to avoid possible issues.

As an example, I disabled multichannel on my hosts. Awful idea but it’s to prove point. And still with our 10.10.0.0/16 subnet set as preferred subnet I ran a backup again.

clip_image018

As you can see the 2*1Gbps LBFO NIC is doing all the lifting on both hosts as it’s switch independent and not LACP load balancing mode we’re limited to 1Gbps.

So how do we control the NICs used with SMB multichannel?

Well SMB Multichannel rules apply. You use your physical design, the capabilities of the NICs and SMB constraints. In reality you’re better off using your design and if needed SMB multichannel constraints to limit SMB to the NICs you want it to use. Do not that disabling SMB Multichannel (client and or server) is a global for the host. Consider this as it affects all NICs on the host, not just the ones you have in mind for your backups. In most cases these NICs will be the same. Messing around with disabling multichannel or NIC capabilities (RSS, RDMA) isn’t a great solution. But it’s good to know the options and behavior.

Some things to note

Realize you don’t even have to set both subnets in the preferred subnets if they are different. SMB kicks of over one, sees it can leverage both and just does so. The only thing you manipulated here SMB multichannel wise is which subnet is used first.

If both of our rNICs would have been on the same subnet you would not even have manipulated this.

Another thing that’s worth pointing out that this doesn’t require your Veeam Backup & Replication VM to have an IP address in any of the SMB multichannel subnets. So as long as the source Hyper-V hosts and the backup target are connected you’re good to go.

Last but not least, and already mentioned in the previous blog post, this also leverages RDMA capabilities when available to help you get the best throughput, lowest latency and leave those CPU cycles for other needs. Scalability baby! No I realize that you might think that the CPU offload benefit is not a huge deal on your Hyper-V host but consider the backup target being hammered by several simultaneous backups. And also consider that some people their virtual machines look like below in regards to CPU usage, in ever more need of more vCPU and CPU time slices.

clip_image020

And that’s what the Hyper-V host looks like during a backup without SMB Direct (with idle VMs mind you).

clip_image022

All I’m saying here is don’t dismiss RDMA too fast, everything you can leverage to help out and that is available for free in the box is worth considering.

Note: I have gotten the feedback that Veeam doesn’t support SMB Direct and that this was confirmed by Veeam Support. Well, Veeam Backup & Replication leverages SMB 3 but that’s an OS feature. Veeam Backup & Replication will work with SMB Multichannel, Direct, Signing, Transparent Failover … It’s out of the Veeam Backup & Replication scope of responsibilities as we have seen here. You feel free to leverage SMB Direct whether that is using iWarp/Roce or Infiniband. This information was confirmed by Veeam and bears the “Anton Gostev seal of approval”. So if SMB Direct cause issues you have a configuration problem with that feature, it’s not Veeam not being able to support it, it doesn’t know or care actually.

Conclusion

The elegance and simplicity of the Veeam Backup & Replication GUI are deceiving. Veeam is extremely powerful and is surprisingly flexible in how you can leverage and configure it. I hope both my previous blog post and this one have given you some food for thought and ideas. There’s more Veeam goodness to come in the coming months when times allows. Many years ago, when SMB 3 was introduced I demonstrated the high availability capabilities this offered for Veeam backups. I’ll be writing about that in another blog post.

Veeam Backup & Replication leverages SMB Multichannel

Introduction

Is it true that Veeam Backup & Replication leverages SMB Multichannel? That is a question that I was asked recently. The answer is yes, when you have a backup design and configuration that allows for this. If that’s the case it will even happen automatically when possible. That’s how SMB 3 works. That means it’s a good idea to pay attention to the network design so that you’re not surprised by the route your backup traffic flows. Mind you that this could be a good surprise, but you might want to plan for it.

I’ll share a quick lab setup where SMB 3 Multichannel kicks in. Please don’t consider this a reference guide for your backup architectural design but as a demo of how SMB multichannel can be leveraged to your advantage.

Proving Veeam Backup & Replication leverages SMB Multichannel

Here’s a figure of a quick lab setup I threw together.

clip_image002

There are a couple of significant things to note here when it comes to the automatic selection of the best possible network path.

SMB 3 Multichannel picks the best solution based on its logic. You can read more about that here. I’ve included the figure with the overview below.

clip_image003

The figure nicely show the capabilities of the NIC situation. To select the best possible network path SMB 3 uses the following logic:

1. RDMA capable NICs (rNICs) are preferred and chosen first. rNICs combine the highest throughput, the lowest latency and bring CPU offloading. on the processor when pushing through large amounts of data.

2. RSS capable NICs: NIcs with Receive Side Scaling (RSS) improve scalability by not being limited to core zero on the server. Configured correctly RSS offers the second-best capabilities.

3. The speed of the NICs is the 3rd evaluation criteria: a 10 Gbps NIC offers way more throughput than a 1 Gbps NIC.

Following this logic it is clear that Multichannel will select our 2 RDMA capable 10Gbps NICs over the management LBFO interface which does not support RDMA and while supporting RSS can only deliver 2Gbps throughput at best. That’s exactly what you see in the screenshot below.

image

Conclusion

So yes, Veeam Backup & Replication leverages SMB Multichannel! Please note that this did not require us to set SMB 3 Multichannel constraints or a preferred network for backups in Veeam Backup & Replication. It’s possible to do so when needed but ideally you design your solution to have no need for this and let automatic detection chose the best network path correctly. This is the case in our little lab setup. The backup traffic flows over 10.1.0.0/16 network even when our Veeam Backup & Replication VM, the Hyper-V host and the backup target have 10.10.0.0/16 as their management subnet. That’s the one they exist on the Active Directory domain they belong to for standard functionality. But as both the source and the target can be reached via 2*10Gbps RDMA capable NICs on the 10.1.0.0/16 subnet SMB3 will select those according to its selection criteria. No intervention needed.

SMB Direct Support

Now that we have shown that Veeam Backup & Replication backups in certain configurations can and will leverage SMB Multichannel to your benefit another question pops up. Can and does Veeam Backup & Replication leverage SMB Direct? The answer to that is also, yes. If SMB Direct is correctly configured on all the hosts and switches their networks paths in between it will. Multichannel is the mechanism used to detect SMB Direct capabilities, so if multichannel works and sees SMB Direct is possible it will leverage that. That’s why when SMB Direct or RDMA is enabled on your NICs it’s important that it is configured correctly throughout the entire network path used. Badly configured SMB Direct leads to very bad experiences.

Now think about that. High throughput, low latency and CPU offloading, minimizing the CPU impact on your Hyper-V hosts, SOFS nodes, S2D nodes and backup targets. Not bad at all, especially not since you’re probably already implementing SMB Direct in many of these deployments. It’s certainly something that could and should be considered when design solutions or optimizing existing ones.

More SMB3 and Windows Server 2016 Goodness

When you put your SMB3 file share continuously available on a Windows 2012 (R2) or Windows Server 2016 cluster (it doesn’t need to be on a CSV disk) you’ll gain high availability trough transparent failover with SMB3 and except for a short pause your backups will keep running even when the backup target node reboots or crashes after the File Server role has failed over. Now, start combing that with ReFSv3 in Windows Server 2016 and the Veeam Backup & Replication v9.5 support of this and you can see a lot of potential here to optimize many aspects of your backup design delivering effective and efficient solutions.

Things to investigate further

One question that pops up in my mind is what happens if we configure a preferred back-up network in Veeam Backup & Replication. Will this affect the operation of SMB multichannel at all? By that I means would enabling a preferred network in Veeam prevent multichannel from using more than one NIC?

I my opinion it should allow for multiple scenarios actually. When you have equally capable NICs that are on different subnets you might want to make sure it uses only one. After all, Veeam uses the subnet to configure a preferred path, or multiple subnets for that matter. Now multichannel will kick in with multiple equally capable NICs whether they are on the same subnet or not and if they are on the same subnet you might want them both to be leveraged even when setting a preferred path in Veeam. Remember that 1 IP / NIC is used to set up an SMB session and then it detects capabilities available, i.e. multiple paths, SMB Direct, RSS, speed, within 1 or across multiple subnets.

I’ll leave the combination of Veeam Backup & Replication and SMB multichannel for a future blog post.

High performance live migration done right means using SMB Direct

I  saw people team two 10GBps NICs for live migration and use TCP/IP. They leveraged LACP for this as per my blog Teamed NIC Live Migrations Between Two Hosts In Windows Server 2012 Do Use All Members . That was a nice post but not a commercial to use it. It was to prove a point that LACP/Static switch dependent teaming did allow for multiple VMs to be live migrated in the same direction between two node. But for speed, max throughput & low CPU usage teaming is not the way to go. This is not needed as you can achieve bandwidth aggregation and redundancy with SMB via Multichannel. This doesn’t require any LACP configuration at all and allows for switch independent aggregation and redundancy. Which is great, as it avoids stacking with switches that don’t do  VLT, MLAG,  …

Even when your team your NICs your better off using SMB. The bandwidth aggregation is often better. But again, you can have that without LACP NIC teaming so why bother? Perhaps one reason, with LACP failover is faster, but that’s of no big concern with live migration.

We’ll do some simple examples to show you why these choices matter. We’ll also demonstrate the importance of an optimize RSS configuration. Do not that the configuration we use here is not a production environment, it’s just a demo to show case results.

But there is yet another benefit to SMB.  SMB Direct.  That provides for maximum throughput, low latency and low CPU usage.

LACP NIC TEAM with 2*10Gbps with TCP

With RSS setting on the inbox default we have problems reaching the best possible throughput (17Gbps). But that’s not all. Look at the CPU at the time of live migration. As you can see it’s pretty taxing on the system at 22%.

image

If we optimize RSS with 8 RSS queues assigned to 8 physical cores per NIC on a different CPU (dual socket, 8 core system) we sometimes get better CPU overhead at +/- 12% but the throughput does not improve much and it’s not very consistent. It can get worse and look more like the above.

image

LACP NIC TEAM with 2*10Gbps with SMB (Multichannel)

With the default RSS Settings we still have problems reaching the best possible throughput but it’s better (19Gbps). CPU wise, it’s pretty taxing on the system at 24%.

image

If we optimize RSS with 8 RSS queues assigned to 8 physical cores per NIC on a different CPU (dual socket, 8 core system) we get better over CPU overhead at +/- 8% but the throughput actually declined (17.5 %). When we run the test again we were back to the results we saw with default RSS settings.

image

Is there any value in using SMB over TCP with LACP for live migration?

Yes there is. Below you see two VMs live migrate, RSS is optimized. One core per VM is used and the throughput isn’t great, is it. Depending on the speed of your CPU you get at best 4.5 to 5Gbps throughput per VM as that 1 core per VM is the limiting factor. Hence see about 9Gbps here, as there’s 2 VMs, each leveraging 1 core.

image

Now look at only one VM with RSS is optimized with SMB over an LACP NIC team. Even 1 large memory VM leverages 8 cores and achieves 19Gbps.

image

What about Switch Independent Teaming?

Ah well that consumes a lot less CPU cycles but it comes at the price of speed. It has less CPU overhead to deal with in regards to LACP. It can only receive on one team member. The good news is that even a single VM can achieve 10Gbps (better than LACP) at lower CPU overhead. With SMB you get better CPU distribution results but as the one member is a bottle neck, not faster. But … why bother when we have …better options!? Read on Smile!

No Teaming – 2*10Gbps with SMB Multichannel, RSS Optimized

We are reaching very good throughput but it’s better (20Gbps) with 8 RSS queues assigned to 8 physical cores. The CPU at the time of live migration is pretty good at 6%-7%.

image

Important: This is what you want to use if you don’t have 10Gbps but you do have 4* 1Gbps NICs for live migration. You can test with compression and LACP teaming if you want/can to see if you get better results. Your mirage may vary Smile. If you have only one 1Gbps NIC => Compression is your sole & only savior.

2*10Gbps with SMB Direct

We’re using perfmon here to see the used bandwidth as RDMA traffic does not show up in Task Manager.

image

We have no problems reaching the best possible throughput but it’s better (20Gbps, line speed). But now look at the CPU during live migration. How do you like them numbers?

Do not buy non RDMA capable NICs or Switches without DCB support!

These are real numbers, the only thing is that the type and quality of the NICs, firmware and drivers used also play a role an can skew the results a bit. The onboard LOM run of the mill NICs aren’t always the best choice. Do note that configuration matters as you have seen. But SMB Direct eats them all for breakfast, no matter what.

Convinced yet? People, one of my core highly valuable skillsets is getting commodity hardware to perform and I tend to give solid advice. You can read all my tips for fast live migrations here in Live Migration Speed Check List – Take It Easy To Speed It Up

Does all of this matter to you? I say yes , it does. It depends on your environment and usage patterns. Maybe you’re totally over provisioned and run only very small workloads in your virtual machines. But it’s save to say that if you want to use your hardware to its full potential under most circumstances you really want to leverage SMB Direct for live migrations. What about that Hyper-V cluster with compute and storage heavy applications, what about SQL Server virtualization? Would you not like to see this picture with SMB RDMA? The Mellanox  RDMA cards are very good value for money. Great 10Gbps switches that support DCB (for PFC/ETS) can be bought a decent prices. You’re missing out and potentially making a huge mistake not leveraging SMB Direct for live migrations and many other workloads. Invest and design your solutions wisely!