As with the Oxygen post, I’ve been playing with the idea and a few Cloudy storage options in the Homelab for a bit. This week, with a well known storage conference going on, I’ve been installing and testing TwinStrata’s CloudArray.

As a side note, when did we stop removing spaces and capitalizing compound words?

About TwinStrata

First things first, you’ll need to know a bit about what TwinStrata really is. Basically, it’s an iSCSI VSA or virtual storage appliance. From their website:

CloudArray® iSCSI storage gateway makes cloud storage as fast, safe and easy to access as local storage. A CloudArray virtual or physical appliance takes minutes to configure and integrates public cloud, private cloud and local or remote storage devices into flexible “Cloud SANs,”

The marketicture diagram:

Downloading TwinStrata

TwinStrata has a number of special offers with various vendors, giving you a free 1TB license of their software backed by various cloud storage providers (you’ll still have to pay for the cloud storage, however). I signed up for and downloaded the Rackspace specific version.

Installing TwinStrata

Side note: I’m beginning to see more and more vendors adopt the vApp/OVA/OVF model for their apps… AWESOME!

To deploy TwinStrata into vSphere, you’ll use the same process you would to deploy any generic OVF. In the vSphere client:

1. File > Deploy OVF Template
2. Next, Next, etc…

Once deployed and booted, you’ll need to browse to the IP that gets assigned:

After a bit of song and dance, you can start to configure:

Fill in the TwinStrata License info from when you downloaded the appliance. Also, configure your admin account:

License agreement… given the options, you know what to do:

Now some of the cool options, compress, dedupe, and encrypt? Yes… I’ll take two. You also have your choice of cloud providers here:

As I said, I chose the Rackspace version, you’ll need to provide this information for your cloud provider of choice:

Note: When it says password here, it means API key

And that’s it. You’re done with this end of things. Next up, using your iSCSI adaptor, add the target, rescan, and enjoy.

Initial Testing

As a quick test, I cold migrated my vMA appliance over to the TwinStrata Datastore. Some things to consider here, TwinStrata is configured with a 5gb disk by default, and will use that to cache things before shipping to the cloud. Once that cache fills… things will get as slow as your connection to the cloud. Here’s a quick look at the graphs from TwinStrata generated by the first sVmotion:

Cold migration to TwinStrata

Booting a VM

Summary:

Overall, performance was not “that bad” considering my bottle-neck is my upstream bandwidth.

Disclaimer: Yes this is a vendor post. No it is not sponsored. Just done using what was available via the TwinStrata website. Also, they may or may not be on the VMware HCL, so… take that into account before going all production on it, etc. tl;dr: It’s not a paid post, also don’t get fired.

Video:

Slides:

For those who haven’t seen his posts / presentations here before should check them out. VMware’s Nathan Small who works as a Staff Engineer at Global Support Services and specializes in… well, storage!

We’ll have him on the 5/2/12 vBrownBag to do what he does best, dive mind-blowingly deep into storage in vSphere environments.

Details

Date/Time: 5/2/12 @ 7P CST
Registration: https://www1.gotomeeting.com/register/613559281

So, this last weekend I’ve decided that I needed to get a bit more secure and a bit more proactive in how data is handled at ProVMware. I’ve been a DropBox customer for years, and likely will stay that way, as it has some great features. However, for everything else, I’ve now got a private install of OxygenCloud.

Why Oxygen?

A few things really, but the biggest driver, is the data lives in a space where I control it. For public access, OxygenCloud tunnels the data back to my client (or from my client back to where the data lives). So, having my data, in a space where I am in control, and access to it is secure… I’m in a better place than I was with DropBox.

Why not Octopus (or similar)?

I’m still playing with options really. When Octopus comes out of beta, you will likely see a post about migrating to it as well. This is one of those spots where any number of tools could best solve the problem, so use the one that fits at a price point that is comfortable.

How To Set Up Private Storage with OxygenCloud

By default, when signing up with OxygenCloud, you get ~5GB space for free:

That first checkbox is the important one. To hook private storage in, I did the following:

1. Built a new Linux VM (Ubuntu 12.4 server) – https://wiki.ubuntu.com/PrecisePangolin/TechnicalOverview/Beta2
2. Added some packages:
   sudo apt-get install libapr1 libaprutil1 lynx build-essential
   (build-essential to get the VMware tools running, lynx you’ll see why in a moment or two)
3. Downloaded the 32bit Oxygen package:
   http://cdn.oxygencloud.com/connector/Oxygen-Storage-Connector_1.0-139_i386.deb
4. Installed it:
   sudo dpkg -i Oxygen-Storage-Connector-1.0-115_i386.deb
5. Started it:
   sudo start oxygenstorageconnector
6. Unlike the beautiful instructions here, I didn’t have a GUI installed, and the configuration service only appears to run on local host during install, so, I fired up lynx:
   lynx http://localhost:18888
7. The rest looks much like the instructions from here.
8. Installed the Mac client
9. Migrated some data & watched it sync.

Caveats / Issues

This process was not entirely seamless… for one, the config service required installing a browser on the server, which works in my lab, but may not work in all environments, so be aware.

Also, the space reporting is a bit off at this stage. My VM has only been assigned 16gb, yet:

The client also seems to find my local storage, but does not seem to connect directly to it while on the same subnet, forcing sync speeds to be slower than I’d like.

Disclaimer

@juliamak of Oxygen promised me a T-Shirt for trying out Oxygen. While T-Shirts are cool and such, I was already in the process of pulling down and evaluating the bits as she made the offer. Thus, free shirts did not influence this post (I don’t think).

We both started and finished vSphere Storage month with Ravi Venkat, and what a wonderfully informative month it was. This weeks recording & slides are below:

Video

Slides

More!

Did you know there are now a metric boatload of ways to get the vBrownBags? In addition to live and the follow-up posts, you can pull the podcast from these places also:

iTunes – http://bit.ly/BrownbagiTunes
RSS – http://bit.ly/BrownbagPodcast

Video

Slides

Video:

Slides:

Find us on iTunes (and other places)

iTunes: bit.ly/BrownbagiTunes
Podcast: bit.ly/BrownbagPodcast
APAC VCAP Brownbag: bit.ly/APACBrownbag

This BrownBag went deep ++ on the fundamentals of storage and how storage is handled in a vSphere environment. Again, sorry for the delay on this one, we had some… technical difficulties.

iTunes:http://t.co/07DLlqC3
Podcast Feed: http://t.co/2FigEC96

Video

Slides

Video

Slides

Notes

• Thin on Thin? Where should you do Thin Provisioning ñ vSphere 4.0 or Array-Level? http://virtualgeek.typepad.com/virtual_geek/2009/04/thin-on-thin-where-should-you-do-thin-provisioning-vsphere-40-or-array-level.html
• Calculating Capacity of a VSA Cluster with RAID10 Configuration http://pubs.vmware.com/vsphere-50/topic/com.vmware.vsa.doc_10/GUID-8941D7EA-A876-4630-A165-80C46D81E219.html
• http://www.yellow-bricks.com/2011/02/18/blocksize-impact/
• vStorage APIs for Array Integration FAQ http://kb.vmware.com/kb/1021976

This time around, it’s audio only, so either view/listen below, or pick us up on iTunes.

Recording

Follow-Up Questions

There is one round of questions that made it’s way on to the web:
http://virtualistmanifesto.com/2011/12/06/storage-smackdown-brownbag-127-a-call-for-questions/

There may yet be more. If anyone feels like transcribing, drop me a line.