Yellow Bricks

by Duncan Epping

Server

#STO7904 VSAN Management Current and Futures by @cdickmann

· ·

Christian Dickmann (VSAN Development Architect) talking about VSAN Management futures in this session, first of all a big fat disclaimer, all of these features may or may not ever make it in to a release and no promises of timelines were made. This session all revolved around VSAN’s mission: Providing Radically Simple HCI with Choice. Keep that in mind when reading the rest of article. Also, this session literally just finished a second ago, I wanted to publish it asap so if there are any typos, my apologies.

First Christian went over the current VSAN Management Experience, discussing the creation of a VSAN Cluster, health monitoring and performance monitoring. VSAN is already dead simple from a storage point of view, but there is room for improvement from an operational point of view, and mostly in the vSphere space. Install / Update / Upgrades of drivers, firmware, ESXi, vCenter etc.

1st demo: HCI Installer

In this demo a deployment of the vCenter Server Appliance is shown. We connect to an ESXi server fist. Then you provide all the normal vCenter Server details like password. Where do you want to deploy the appliance? How about on VSAN? Well you can actually create the VSAN Datastore during the deployment of the VCSA. You specify VSAN details and go ahead. During the install/configuration process VSAN will simply be configured using a single host cluster. When vCenter is installed and configured you simply add the rest of the hosts to the cluster. Very cool if you ask me!

2nd demo: Simple VMkernel interface creation

In this demo the creation of VMkernel interfaces is shown. Creation of the interfaces is dead simple as you can simply specify the IP ranges and it does this for every host using the specified details. Literally 4 hosts and interfaces were creates in seconds.

3rd demo: Firmware Upgrade

In this demo in the VSAN Healthcheck it is shown that the firmware of the disk controller is out of date. When you say update, vendor specific tools are downloaded and installed first. When this is completed you can remediate your cluster and install drivers and firmware for all nodes in your cluster, all done through the UI (webclient) and literally in minutes in a rolling fashion. I wish I had this when I had to upgrade my lab in the past.

4th demo: VUM Integration

80% of vSphere customers use VUM so integrating VSAN upgrades and updates with VUM makes a lot of sense. During the upgrade process VUM will validate which version of vSphere/VSAN is supported for your environment. If for whatever reason the latest version is not supported for your configuration it will make a recommendation to use a different version. When you remediate VSAN provides the image needed and there is no need even to create baselines etc. All of this manual work is done by VSAN for you. Upgrades literally become 1 or 2 clicks, and all risks are mitigated by validation of hardware/software against the compatibility matrix.

5th demo: Automation

In this demo Christian showed how to automate the deployment of 10 ROBO clusters end to end using PowerCLI. One by one all the different locations are being created. Every single aspect is fully automated, including even the deployment of the witness appliance. The second demo was the upgrade of the VSAN on-disk format using python. In a fully automated fashion all clusters are upgraded in a rolling fashing. No magic here, all using public APIs.

6th demo: VSAN Analytics

Apparently with the 6.2 Christian found out that Admin’s don’t read all KB articles VMware releases, based on the issue experienced with a disk controller he decided to solve this problem. Can we pro-actively inform you about this problem? Yes we can, using a “cloud connected” VSAN Healthcheck we know what you are using and we can inform you about KBs and potential issues and recommendations that may apply to you. And that is what was shown in this demo, a known issue is bubbled up through the healthcheck and the KB details are provided. Mitigating is simply a matter of applying the recommendation. This is still a manual step, and probably will stay as Christian emphasized as you as the administrator need to have control and should make the decision whether you want to apply the steps/patches or not.

Concluding, in literally 40 minutes Christian showed how the VSAN team is planning on simplifying your life. Not just from a storage perspective, but for your complete vSphere infrastructure. I am hoping I can share the demos at some point in the future as they are worth watching. Thanks Christian for sharing, great job!

@DuncanYB’s recommended reads part 2

· ·

The last edition was 2 weeks ago, in that article I pointed to the rumour of Pernix being acquired by Nutanix and since then it has gone silent. Well almost silent, Frank Denneman has announced he is no longer with Pernix (which was picked up by The Register) so I think it is safe to say there is a truth to the rumour. Sad moment as I really liked the company and really liked the technology, hope the product will live on. I guess time will tell and I am sure that by the time the next edition of recommended reads is published we will have seen an announcement.

  • Springpath to focus on Cisco OEM Development
    It is an interesting development if true. Springpath is one of the “newer” hyper-converged players and initially sold through various OEMs. Now they seem to solely focus on Cisco, which I guess makes sense as it will make it easier to work on deeper integration with Cisco tooling. Wouldn’t surprise me if Cisco picks these guys up at some point in time.
  • Build a Xeon D-1500 (Open) Home Lab with me? by Joep Piscaer
    I’ve had this Intel proc family and SuperMicro board on my list for the longest time for a homelab, never got around to ordering anything but I have to agree it is a great piece of kit with a lot of power combined with relative low power consumption and a great WAF (wife acceptance factor), read Joep’s post and his wiki entry on OpenHomeLab.org for more details
  • On the topic of homelabs: Supermicro SuperServer SYS-E300-8D and SYS-E200-8D close-up pictures by Paul Braren (another one by Paul here on the same subject)
    This is a nice piece of kit. It also holds the D-1500 CPU, is nice and small and has 10GbE on-board. Could be great for a VSAN lab, small but still packing a lot of power.
  • Thoughts on Pokemon G0
    Nothing to do directly with Infrastructure, but I had to list this as I think it is a brilliant concept. Especially the use of the camera is smart, and I bet it got many kids walking around the neighbourhood instead of sitting on the couch playing on their PS4. Another interesting aspect is that it seems to be bringing people together and have people talking whom normally probably wouldn’t be talking to each other. Plus of course create traffic chaos and dangerous situations as no one is paying attention anymore where they are walking / driving.
  • Disrupt Your Data Center with Transformative CI & HCI Technologies by Jason Nash
    Well not really a blog, but an awesome panel session with Jason Nash, Chad Sakac, Dheeraj Pandey, and Matt Smorto on the hyper-converged market today. Some great insights and good discussion. Always impressed with Chad, I just love the way he handles the discussion/situation during these types of conversations.

    • Media Headlines by Chad Sakac. Setting the record straight when it comes to the Dell/EMC strategy and hyper-converged offerings. This was a response to an article in The Register, not to the video by the way.
  • Samsung 850 EVO SSD 4Tb Review
    Not the most exciting article, but more important here is 4TB SSD which is relatively affordable. Prices are still dropping, can I have that 16TB SSD now? Please?!
  • How the queen of Silicon Valley is helping Google go after Amazon’s most profitable business
    Some interesting bits and pieces in here, especially how the organization is being transformed by Diane within Google.
  • NUMA Deep Dive Part 1: From UMA to NUMA by Frank Denneman
    Last one on the list, so it has to be something good. You like deepdives? Well this is going to be your cup of tea then. Frank is diving in to it deep immediately and even followed up with Part 2 in the same week. And this week part 3 and 4 also dropped. Those diagrams by the way are insane… Considering Frank is looking for a new challenge, I wouldn’t be surprised if Intel made him an offer.

VMworld schedule builder live…

· ·

For those interest, the VMworld session schedule builder is live now and it is part of the content catalog, just log in and start registering. I created a top 15 yesterday, make sure to read that for some tips. Note that sessions tend up to fill quickly, so don’t wait to long with creating a schedule.

Oh, and this nice teaser was created by the VMworld team based on the Quick Talk on VSAN Use Cases we did last year (Lee Dilworth and I). Watch it 🙂

Another major milestone, 5000 customers reached for VSAN!

· ·

Yes, new quarterly results are out, and this is crazy, Yanbing (SVP/GM for Storage & Availability BU) dropped the news on twitter last night:

Yes that is correct, VSAN customer count grew with 1500 in just 1 quarter and the license bookings also grew > 200% Y0Y. Crazy numbers if you ask me, and some serious traction. Considering how well VxRail is doing I can only see this ramping up faster even. Just some quotes from last nights earnings call, you can find a transcript here and audio here if you are interested.

Pat Gelsinger:
Our hyper-converged license bookings, including VSAN and VxRail software, grew over 200% year-over-year. The new VxRail appliance we jointly developed with EMC was launched earlier this year and has enjoyed an impressive start, benefiting also from Dell’s decision to resell VxRail.

The rock stars of Q2 were NSX and VSAN and they were just fabulous

Yes, clearly the VxRail and the momentum that it is seeing was a piece of that. But a lot of the customer count was non-VxRail. So this product has now hit a point of inflection, as you suggest, that really was thrilling. And if I would call out one product in Q2 that just blew their numbers away, VSAN is it, and it was the rockstar of Q2 above all others. And what we’re seeing is that the product maturity, the channel momentum, I commented on the transactional channel before, and that really is well fit to VSAN. But also some of the enterprise use cases have clearly seen momentum, as well. And then we add to that the incremental benefits of VxRail that we jointly developed with EMC and Dell now reselling that VxRail product. It was also geographically balanced, as well. We saw VSAN growing across all of the geos, large customers, major wins, in healthcare, also mid market wins across different geos. So really, really excited about the VxRail, as well as the overall hyper-converged category for us and VSAN, the rockstar of Q2.

You know what, I will give away 5 “Essential Virtual SAN” books to 5 random people who leave a comment (and retweet / share this post on linkedin) to celebrate this great achievement. Please leave comment with your real name and correct email address so I can reply for your home address! I email winners Wednesday 19th of July end of day. [UPDATE: Winners have received an email from me, or will within the next 10 minutes]

How thermal paste can impact VM performance

· ·

On twitter a tweet from Frank flew by pointing to an article which was written by one of my VMware colleagues: Matt Bradford aka @VMSpot. I hadn’t seen the article, while it was written in 2014 and I am surprised it never caught more attention. Matt describes in his post how the use and placement of thermal paste can influence VM performance. Who would have thought of that, and I am seriously impressed they managed to get to the bottom of this!

We haven’t had our HP BL460c Gen8’s with the new Xeon E5-2697 v2 12 core processors long. Last week we started to get e-mails from the help desk that users were complaining about sluggish performance in Citrix. Oddly, all of the XenApp VM’s happened to live on the same ESXi host. I say oddly because performance issues rarely seem to fall in line as they did here. We immediately evacuated the host and admitted it to the infirmary cluster.

…..

It didn’t seem to matter if the CPU’s were under load or idle, the temperature would not stray from 69°c. This had to be an issue with the temperature sensors, I thought. So we pulled the host and removed the heat sinks so we could look at the CPU’s through a thermal camera we borrowed from engineering.

I am not going to post the full article here, go over to Matt’s blog and have a read. It is flabbergasting if you ask me, and definitely one of the coolest reads in a long time. And thanks Frank for bringing this one up. I just had to share in on a broader platform.

That reminds me, maybe it is time to bring back my “favourite reads” post I did for a long time on the VMTN Blog, but host it here instead. Hmmm. Ah well, lets make a start here and follow up with “Recommended reads” posts in the future:

That’s it for now, enjoy reading.