Yellow Bricks

by Duncan Epping

Server

Extending your vSphere platform with Virtual SAN

· ·

Over the last couple of months I’ve spoken to many customers about Virtual SAN. What struck me during these conversations is how these customers spoke about Virtual SAN. In all cases when we start the conversation it starts with a conversation about what their environment used to looked like. What kind of storage they had. How it was configured, number of disks etc you name it. Of course we would discuss what kind of challenges they had with their legacy environment. Thinking back to these conversations there is one thing that really stood out, although never explicitly mentioned, the big difference between Virtual SAN and traditional storage systems is that Virtual SAN is not a storage system but rather an extension of the VMware vSphere Platform.

Source: Wiki
Software extension, a file containing programming that serves to extend the capabilities of or data available to a more basic program

I believe this statement is spot on. What is great about Virtual SAN is that it does the extension of the capabilities of vSphere in an extremely easy way. Virtual SAN achieves this simply by abstracting layers of complexity and pooling the resources and allow these to be assigned to workloads in an automated fashion whether through the use of policies and a simple UI or through the vSphere APIs. Keywords here are definitely: abstract, pool and automate.

Maybe I should have used the word “converging” instead of “abstracting”. That is essentially what is happening, and although many other vendors claim the same, I truly believe that Virtual SAN is one of the few solutions which is truly hyper-converged as it seamlessly converges layers instead of adding a layer on top of another layer. Hyper-convergence is more than just stacking layers in a single box.

With Virtual SAN storage is just there. Not bolted on, layered on top or mounted to the side, an integral part of your environment, an extension of your platform. Virtual SAN does for storage what vSphere does for CPU and Memory, it becomes a fundamental component of your cluster.

More than 8 disks in your Virtual SAN host?

· ·

It seems Virtual SAN is on fire… I’ve been getting emails about VSAN configurations on a daily basis from partners, VMware field and customers. More and more customers seem to be looking at deploying Virtual SAN with a high number of disks and this week I ran in to something that I wanted to share with you. I thought it was common knowledge, but I guess I was wrong. If you are planning to have more than 8 disks / devices in your Virtual SAN host then it is good to know that you will either need a second disk controller (Check the HCL!) or use a SAS Expander! This design / sizing principle is listed in the VSAN Ready Node PDF, and should also make its way to the Design / Sizing Guide soon. The following is mentioned in the VSAN Ready Node PDF:

No more than 8 disks (SSD + HDD) supported behind a single controller. In case you need to add more than 8 disks, please add an additional controller.

I met up with the product manager for the compatibility guide and asked him which SAS Expanders are supported, he mentioned that support is done per platform and that the following vendors and platforms have just been update. Note that the HCL will still need to be updated for these:

  • Dell SAS Expanders 5.5 and 6.0 hybrid. Supported on a specific configuration: H730 on 730d platform with 24 drives.
  • Lenovo SAS Expanders 5.5. Supported on a specific configuration: Lenovo 720ix, RD650 platform with 12 drives.

We also are working on support for HP and Cisco (and others over time), and you can also expect an update on those soon. For now, keep in mind that when designing / planning for Virtual SAN the number of virtual disks may mean you need to change the configuration of your host.

Win a VMworld Party VIP ticket!

· ·

Are you attending VMworld and are you a big fan of Neon Trees and/or Alabama Shakes? The VMworld team was kindly enough to provide me with a VIP ticket with the opportunity to meet Neon Trees and Alabama Shakes at the VMworld Party in the AT&T Park. Awesome location and great bands. I was thinking about it, and I figured that one of my readers / followers must be a huge fan of both bands or either of the bands and would go wild even thinking about the opportunity to meet them, so why not give the tickets away! So here is your chance to win a “meet and greet” with the band. Here’s how to win and some rules as I want to make sure it is a customer or partner winning this:

  1. You need to be attending VMworld, with a ticket valid to enter the party (paid attendee)
  2. Leave a comment and explain to me how big of a fan you are of either of the two bands and why YOU deserve to meet your heroes!
  3. Make sure you leave a valid email address in the email field!

I will select a winner on the 1st of August.

High latency VPLEX configuration and vMotion optimization

· ·

This week someone asked me about an advanced setting to optimize vMotion for VPLEX configurations. This person referred to the vSphere 5.5 Performance Best Practices paper and more explicitly the following section:

Add the VMX option (extension.converttonew = “FALSE”) to virtual machine’s .vmx files. This option optimizes the opening of virtual disks during virtual machine power-on and thereby reduces switch-over time during vMotion. While this option can also be used in other situations, it is particularly helpful on VPLEX Metro deployments.

I had personally never heard of this advanced setting and I did some searches both internally and externally and couldn’t find any references other than in the vSphere 5.5 Performance paper. Strange, as you could expect with a generic recommendation like the above that it would be mentioned at least in 1 or 2 other spots. I reached out to one of the vMotion engineers and after going back and forth I figured out what the setting is for and when it should be used.

During testing with VPLEX and VMs using dozens of VMDKs in a “high latency” situation it could take longer than expected before the switchover between hosts had happened. First of all, when I say “high latency” we are talking about close to the max tolerated for VPLEX which is around 10ms RTT. When “extension.converttonew” is used the amount of IO needed during the switchover is limited, and when each IO takes 10ms you can imagine that has a direct impact on the time it takes to switchover. Of course these enhancements where also tested in scenarios where there wasn’t high latency, or a low number of disks were used, and in those cases the benefits of the enhancements were negligible and the operation overhead of configuring this setting did not weigh up against the benefits.

So to be clear, this setting should only be used in scenarios where high latency and a high number of virtual disks results in a long switchover time during migrations of VMs between hosts in a vMSC/VPLEX configuration. I hope that helps.

Shift in focus… Go Storage & Availability OCTO!

· ·

Almost a year ago I joined the Office of CTO under Paul Strong. My main focus was SDDC, but I naturally gravitated towards the core platform (vSphere) and software defined storage and topics like availability. Not just my personal preference, but also a common requested topic for public speaking engagements. Most VMUG speaking requests I receive are around VSAN, VVols or vSphere HA. Each year I take some time to reflect on where I am, what I do, and where I want to go. This year I asked myself what really excited me in todays world of IT/infrastructure? What am I most passionate about? What do I enjoy talking and writing about the most?

Having written books on Virtual SAN and vSphere Clustering, and countless blog posts on the topic of software defined storage, BC/DR and availability it was pretty obvious what I am most passionate about. I like talking and writing about Virtual SAN, Virtual Volumes, Site Recovery Manager and it is safe to say that I am a vSphere HA fanboy. I am most passionate about Storage & Availability, that much was obvious

At an internal event I had a conversation with Charles Fan and Christos Karamanolis. The Storage & Availability BU was considering creating an Office of CTO and they asked if I would be interested in collaborating in some shape or form. For me this was a no-brainer. Knowing what is coming for Virtual SAN and Virtual Volumes (and future products we are working on) I asked myself if collaborating would be the best option or if I should take that next step. The decision was easy, as of this week I have officially joined the Office of CTO of the Storage & Availability BU.

In the Office of CTO I will be responsible for connecting our R&D team with customers, partners and our field. I will be evangelizing software defined storage and availability, primarily in EMEA and APJ. I will focus on defining and communicating VMware’s vision and strategy, and be an active advisor for our product roadmap and portfolio. I can’t be more excited than this, I am super enthusiastic about all what is to come out of our business unit and it is extremely energizing to say the least to talk to our customers about what we do today and what is coming tomorrow. As a big plus I get to work with my friend Rawlinson Rivera once again, and report in to someone I greatly respect namely Christos who will be heading up the team. Make sure to read Christos’s blog post on the team that has being formed and some hints of what you can expect in the future. Lets get busy!

Thanks Charles, Christos and Paul for this great opportunity!