In our Twitter vSAN Chat with Duncan Epping and Cormac Hogan, we covered the A to Z of their newly published “vSAN 6.7 Deep Dive” book. If you’ve ever wondered “What exactly is vSAN?”, everything to know about Stretched Clusters, or the vSAN 6.7 Deep Dive dedication song “Turn the Page” written by Bob Seger and covered by Metallica, this Tweet Chat covered it.
We hope you enjoy this Q&A and we invite you to join us at our next Tweet Chat on January 31st at 11am PT with the vSAN and vSphere teams!
#vSANChat Introduction: Can you tell us who you are and where you’re joining us from? (tweet)
My name is Duncan Epping and I work in the vSAN Business Unit within VMware as a Chief Technologist… and I live in The Netherlands, also known as “Holland”.
Cormac here from VMware Storage and Availability BU – Office of the CTO. I’m based in Ireland and looking forward to our #vSANchat!
Q1: First question to get us started… what exactly is vSAN and how has it evolved over the years and in the latest release? (tweet)
Duncan: vSAN is a software-based distributed storage solution which tightly integrates with VMware vSphere and is what we call “policy driven”. Which means that you as the admin define the requirements of the Application or VM in a policy and vSAN places the data based on that!
We have gone from basic “VM storage” for “hybrid” only to a hyperconverged platform that can deliver storage for any app and any use case, even across different locations etc.
Cormac: To being with, vSAN is VMware’s Hyperconverged Infrastructure offering, bringing storage and compute together as a single offering.
It has evolved considerably over the years, supporting newer flash technologies and adding more and more data services (erasure coding, dedupe, compression, encryption, etc.) over the years.
Q2: What are the storage pain points you hear most from customers, and how can vSAN help to solve them? (tweet)
Duncan: In most cases customers are looking for simplicity, availability and consistency. vSAN easily hits all three of them. Much easier to manage than a dedicated storage platform, great consistent performance and on top excellent availability.
Q3: How can we gain maximum value from vSAN? What are the tips and tricks from the frontlines? (tweet)
Duncan: it is all about understanding the power of policy-based management and the operational model. Yes, there still is an operational aspect to take in to account!
Q4: A main focus of the book is simplicity – why is this so important? (tweet)
Duncan: Because in this day and age most customers can’t afford to spend countless of hours managing a system which has no direct business value. They are looking for simple/easy solutions that allow them to focus on applications/services that deliver value to the company.
Cormac: Reiterating what Duncan said, IT can be complex, and in a lot of cases it is unnecessarily so. While vSAN can be complex under the covers, it should not be complex from a deployment and management perspective. That’s what the book is highlighting.
Q5: If there is one thing you want us to gain from reading the book, what would it be? Why? (tweet)
Duncan: the key thing in my opinion is that it is all about Policy Based Management and the simplicity that this brings, and the risk that it removes. When using policies customers typically make no (or far less) placement mistakes as a result.
Cormac: I would say that we want to enforce the idea of policy driven storage. You no longer deal with carving up hardware to meet a specific requirement. In vSAN, you simply create a policy with your requirements and ask vSAN to take care of it.
Q6: A whole chapter is devoted to a very specific type of vSAN configuration – stretched clusters. What are the key considerations and learnings from Stretched Cluster use cases? (tweet)
Duncan: The key thing is the ease of implementation and flexibility it provides! vSAN Stretched Clusters are configured in minutes not weeks, and the policy allows you to mix and match “local” and “stretched” with various protection mechanisms in a very easy way!
Q7: You dedicate the book to the song ‘Turn the page’ by Metallica/Bob Seger. What’s the story there? (tweet)
Cormac: A7. Apparently, since I am the older of the two of us, I am supposed to answer this one. 🙂 There are many SABU team members who travel and spend time away from home delivering the vSAN message. This is a nod to those guys and girls.
Also, ‘Turn the page’ is a nice message to include in a book dedication
Duncan: For me it was definitely the lyrics that did it, those just align well with what Cormac and I have experienced in the past years. Going from country to country, event to event endlessly. Many vSAN Road Warriors out who deserve a pat on the back!
Q8: Let’s open it up! What questions do you have for Duncan and Cormac today? The sky’s the limit!
Q: Is there a requirement for using brocade switches? (tweet)
Duncan: No there is no requirement for that. We have customers using different types of switches ranging from Cisco to Arista and Brocade indeed. But it is up to you as a customer!
Q: Why is dedup and compression not an option across disk groups? (tweet)
Cormac: It was a design decision at the time – dedupe and compression is done on a per disk group basis at present. Of course, this ‘may’ change in future releases.
Q: In vSAN, are compressed data blocks that need to be moved to another diskgroup (rebalance) rehydrated and only compressed again in the new diskgroup, or are compressed blocks transferred? (tweet)
Cormac: Since we do dedupe and compression together, they would need to be rehydrated and un-deduped (is that a word?), then moved to a new diskgroup where the dedupe AND compression would need to take place again.
Q: How does the price compare to the market leaders for SAN? (tweet)
Duncan: that is a challenging question as it will depend on configuration/sizing/licensing. In most cases however we see vSAN being substantially cheaper than traditional storage systems. With the added benefit that there’s no huge upfront cost, pay as you scale/grow!
Cormac: Difficult to say as it is completely dependent on the components you decide to use for vSAN (flash devices, etc). The advantage though is that you can start small (2-node, 3-node) and grow as you need.
Q: What is the difference between he different license versions? (tweet)
Duncan: Licensing details can be found in this great white paper: http://bit.ly/2LZskA8.
Q: How does the performance of vSAN compare to traditional enterprise SAN solutions? (tweet)
Duncan: That is difficult to say as it very much depends on the configuration of both vSAN and the tradition array, in most cases though I would say that vSAN performs a lot better, primarily as it was developed with flash and virtualization in the back of our minds.
Q: How does the performance of vSAN compare to traditional enterprise SAN systems? When I chat to our storage team they seem to believe it can’t compete so would love to prove them wrong. We recently installed our first vSAN cluster so it’s still very new for us.
Cormac: There are a number of performance papers out there which talk about how well certain applications run on vSAN. But this is very much dependent on the vSAN configuration, network bandwidth between hosts and speed of flash devices used. It compares very well (tweet)
Q: I‘m curious if different vSAN datastores in one cluster (different disk groups per host joined in different vSAN datastores) will be possible? Like one all-flash and one hybrid tier?
Duncan: Always difficult to answer roadmap questions as you know! Yes, we have had this request before but we also see the world moving more and more towards all-flash. Although not in all part of the world is flash affordable, we are getting there! (tweet)
Q: Is it possible to migrate a vSAN 6.6 cluster running on VC 6.5 appliance to a new 6.7 VC and then upgrade to vSAN 6.7. We are using Distributed switches.
Cormac: Yes this is possible. There are a bunch of upgrade consideration papers available on Storage Hub– please review beforehand (tweet)
Q: I’m wondering if there is or will be an option to choose how aggressively vSAN marks a disk or disk group as failed and then removes it from vSAN to prevent any impact. (tweet)
Cormac: Thanks for the question. We have a feature called DDH – Degraded Disk Handling – which is doing precisely that. It has gone through a number of iterations and updates in various vSAN versions but should do exactly what you are requesting
Q: Could you maybe already share, if local redundancy (between disk groups on same host) is on the roadmap for vSAN (new feature of HyperV 2019) (tweet)
Duncan: Are you trying to get us fired? Yes it is definitely on our radar. I can’t discuss any timelines unfortunately as you can imagine.
You can always ask a local VMware connection for a roadmap session when needed, they can give ALL the details under NDA.
Q: My production servers are running vSAN 6.6. Should I update them right now to vSAN 6.7 U1 (or should I wait a little bit)? (tweet)
Duncan: Depends on how comfortable you are. Personally, we have been using 6.7 U1 fine and we had great feedback from customers so far. But in some cases, for instance your backup vendor may not support it yet. So, check all dependencies!
Cormac: We are always improving with each release. Big difference between 6.6 and 6.7U1 are around resync improvements as well as some guard rails around maintenance modes, etc. Of course, we also introduced TRIM/UNMAP for reclaiming space, so well worth considering (tweet)
Q: I‘m curious if different vSAN datastores in one cluster (different disk groups per host joined in different vSAN datastores) will be possible? Like one all-flash and one hybrid tier?
Well, let‘s say, two different disk groups per host (1x all-flash / 1x hybrid) and other all hosts joined in different vSAN datastore tiers. (tweet)
Cormac: As far as I know there are no plans for configurations like this.
Q: Backup is a good topic! Not only check the HCL but also look with your backup vendor if the specific version is supported before you going for an upgrade. (tweet)
Cormac: Actually, the later releases of VUM has a whole bunch of pre-checks which will handle this for you going forward. But yes, always backup!
Q: We have a customer using vSAN ROBO because of the small many remote offices they use, works well due to hardware cost and implementation compared to other SANs. (tweet)
Duncan: That is great to hear. Thanks for the confidence!
Thanks for joining us! Follow the hashtag #vSANchat on Twitter for the next chat!
