Thin Provisioning and Fragmentation

By On · Add Comment

Frag-ThinAfter a recent post summarizing my thoughts around on thin-on-thin provisioning there were a few comments on the Twitterverse (shout outs to @trodemaster) expressing concerns about VMDK fragmentation by using thin provisioned virtual machines (on thin provision LUNs). This is yet another topic that has been heavily discussed, but here we go again. The thing about fragmentation is that it happens no matter what when you use thin-provisioned VMDKs. Whether you configure thin-on-thin provisioning, or thin-on-thick provisioning the virtual machines are going to need space allocated when write commands are issued.

The use of thin provisioning, then, boils down to the answer to one very important question; Is the virtual machine an appropriate candidate for thin provisioning?  Successful use of thin provisioning is predicated on choosing the right virtual machines/application to provision in thin format. Obviously, if the application running in a thin provisioned virtual machine is a write intensive type of application and its rapidly growing and allocating new blocks (we call that a fatty), then this is not the type of application you wan’t to deploy in a thin provision format.

Thanks to Storage vMotion we can remedy these situations by changing the format from thin to thick.  Remember too, you have the ability to define the thin or thick format per disk, on multi-disk virtual machines you can chose thin for some disks, like the operating system disk, and then use thick for other disks, like the data disk. Fragmentation can and will happen when using thin provisioned VMs, but whether or not it poses a problem lies in  knowing when to use it and when not to.

- Enjoy.

For future updates, be sure to follow me on Twitter at  @PunchingClouds

 

Getting Thin (on Thin) “It’s your Responsibility”

By On · 4 Comments

ThickThinRecently while at EMC World I’ve been feeling more like a P90X fitness coach (I wonder who they have me confused with) than a technologist with the number of question I’ve been getting from customers regarding Thin provisioning. Many customers have mentioned to me how the storage vendors recommend avoiding thin provision virtual machines an top of thin provision LUNs. After doing some cursory research on the subject its seams the reason behind this is driven more by administrative operation than technical disadvantages. Let the record clearly show that I do agree with their assessment for the administrative danger of doing this. However I believe with proper monitoring and response planning those dangers can be mitigated.

In order to avoid the dangers of  thin-on-thin as suggested by the storage vendors you can choose between “thick-on-thin” and “thin-on-thick”. Thick-on-Thin means thick virtual machines on thin provisioned LUNs and Thin-on-Thick will in either case you’re only concern with running out of space on of the two elements but not both.

The issue with thick on thin arrises when multiple virtual machines are created in a short period of time against the same datastore built of a thin provision LUN for example if I crate an 800GB thin provision LUN and within the first 2 days of creating the LUN I deployed 7 new virtual machine each with 100 GB VMDK I now consumed all but 100 Gigs of the LUN. In fact I probably consumed more when you take into account all of the additional files required to the virtual machines. So was it really worth thin provisioning the LUN if my attempt at achieving space efficiency lasted less than 2 days.

Barring any technical limitation that I didn’t find, thin-on-thin provisioning is going to give efficient utilization of space at both they hypervizor and the array level. this dual efficiency, however, comes at the expense of the administrative effort involved in establishing the proper monitoring, alerting, and response action planning. It’s not enough to just know of the impending danger of space consumption you must know how to react by establishing what actions will be taken when alert thresholds have been broken

Although this blog doesn’t come with any 90 days guaranteed or offers for protein shake powders, I do hope you’ve I have provided you with tools to help you get thin (on thin). Feel free to take and send before and after photos of your transformation.  : )

- Enjoy

For future updates, be sure to follow me on Twitter at  @PunchingClouds

vOpenData and Storage Dashboard

By On · Add Comment

vOpen-LogoBy now a great deal of folks in the community have heard and know about vOpenData and what it is, who is behind this cool project and what it does. For those of you who don’t know what vOpenData is or what it does vOpenData is an open community project that came to life based on user requests for knowledge about virtualization environments and their most common configuration ranging from average VMDK size, number of virtual machine deployments, etc. VMware’s own and my boy! (You ma Boy Blue) William Lam is one of the minds behind this useful can cool project.

By leveraging the volunteered community data and apply simplified analytics vOpenData can provide answers to a great deal of questions that people may be interested in acquiring but yet it may be difficult to get.

vOpenData collects various types of data from virtualized infrastructures with an effort to prevent the collection unique data points that could compromise infrastructure private information such as hostsnames, IPs, UUID, etc.

The data collected is used to virtualization based statistics and data modeling for the community that can be useful in a great deal of scenarios such as average VMDK sizes, number of virtual machines, most utilized or deployed hardware, average hosts connected per vCenter servers, etc.

vOpenDate Infrastructure Dashboard

vOpen-Infra

By now, the entire community has shown interest on the vOpenData project and all the blogs have been raving about it from the very first day they saw it. Because of that, I didn’t add my two cents and wrote something about. I figured the project had gained enough attention from the blogger community and just waited for thing I can mention that was particularly of interest to me.

New to vOpenData statistics is the collection of data with regards to storage arrays. Now vOpenData can produce statistics gathered from storage system, LUN sizing, arrays by vendor types and more. As storage is one of the last pain points to be simplified, the more knowledge you have of what is out there and what is being use and in what capacity can help in many ways.

vOpenData Storage Related Dashboard

vOpen-Storage

I’m a big fan of this new storage statistic capability and of vOpenData over all. This will help me with some of my queries with regards to storage centric topics in conjunction with my current responsibilities at VMware. I would highly recommend to give vOpenData project a look if you haven’t done so yet and sign up if possible and contribute to the communities wealth of valuable information and statistics.

- Enjoy

For future updates, be sure to follow me on Twitter at  @PunchingClouds

VSA Upgrade Required after Applying vCenter Server 5.1U1 Patch

By On · Add Comment

VSA-Video.pngHeads up folks! for those of you that are using the vSphere Storage Appliance 5.1.1 (VSA). If you have or are planning to apply the vCenter Server 5.1U1 patch, you must also upgrade the Sphere Storage Appliance (VSA) to version 5.1.3 after deploying the vCenter Server 5.1U1 patch due to the fact that the Sphere Storage Appliance 5.1.1 (VSA) is not compatible with the latest vCenter Server patch release.

Some of the changes added to the new release are:

  • Increase of the VMFS heap size to 256MB by default. This change will allow ESXi hosts to address up 24TB of storage capacity.
  • Support for 10 Gigabit Ethernet NICs

Check out the vSphere Storage Appliance 5.1.3 (VSA) release notes for more details on the fixes and features of this release.

- Enjoy

For future updates, be sure to follow me on Twitter at  @PunchingClouds

Austin, Texas VMUG

By On · 3 Comments

VMUG-Logo.pngAlright!!!! VMware community members and followers, If you’re anywhere near Austin, Texas on April 2nd come an join us for a day of fun and knowledge transfer at the VMware User Group (VMUG). The event is set to be a very good one with great breakout sessions about the Software Defined Datacenter (SDDC), Software Defined Storage, VCDX Workshop bootcamps.

If you interested on the challenge and pursuing the VCDX certification you want to attend one of the bootcamps delivered by actual VCDX panel members. The VCDX bootcamps have shown to be very helpful for candidates. I will be there along other fellow VCDXs and  panel members Wade Holmes, and Matthew Meyer.

The VCDX bootcamp will cover everything you need to prepare for the certification process and the defense with topics such as

  • What the VCDX covers, and what it is designed to demonstrate
  • Profiles of successful and unsuccessful candidates
  • An insider’s view of the VCDX Panel Defense process, including perspectives from the panelists
  • Advanced best practices for VMware design and architecture available nowhere else
  • Practice VCDX Design and Troubleshooting Scenarios, and more

If you’re curious and want to learn more about the Software Defined Datacenter and Software Defined Storage, I will also be presenting a session on VMware’s Software Defined Storage (SDS) vision for the Software Defined Datacenter, and providing some insight into some of the features and technologies VMware is currently working on to make the vision a reality.

If you are already a VCDX, master of the universe and already know about all this stuff than great!, you’re the Man!!! or the Woman!!! still just come down and socialize with everyone it will be fun.

Feel free to sign up and register Austin VMUG User Conference

- Enjoy

For future updates, be sure to follow me on Twitter at  @PunchingClouds