Storagezilla

Kill your Blackberry

Doesn’t have to be a Blackberry, might be an Android or iOS Device, but I’ve been without mobile email since November when airport security broke my Blackberry. At the time I had no idea who I’d go to replace it, so I didn’t. And to be quite direct it’s been a liberating experience.

I haven’t given up on email. I actually like email as a communications medium, it allows me to take a moment or more than a few moments to decided the correct course of action and doesn’t require a ‘don’t-confuse-me-with-the-facts’ snap response some phone callers look for. But I just might have given up on email on the go.

These days I actually read email instead of skimming it on a small screen as quickly as possible and missing things because I want to get back to my life.

I’m present at the dinner table. There’s no siren song from a piece of plastic with a flashing notification light anywhere in the room or in the car.

I no longer wonder if anything important has happened which I need to deal with the moment they open the doors of the plane I’m on or after a movie or some other social engagement.

Everything can wait until the next time I log in.

In meetings unless I have my laptop open my options are to either listen and interact or lean back and yawn as loudly as possible to let the room know they’re boring me senseless with their same old palaver.

While the people sending me email haven’t changed I don’t find myself annoyed going into the weekend anymore. I suspect mobile email has a magnification effect, if you’re out living your life and you get something unpleasant into your inbox it appears to be far worse then & there then it would if you were in work mode in front of a computer screen.

All in all I’m coming up on six months without mobile email and the universe hasn’t collapsed.

Who’d have thought it?

EMC Loves Stat Geeks, “Oracle has the Most to Lose” - Forbes

“We see an emerging role around data scientists” inside large corporations, says Jeremy Burton, EMC’s Chief Marketing Officer. “If we can build influence there, intellectual capital there, we’ll do well.” EMC is holding its first-ever meeting of industrial statisticians in Las Vegas this May, he says, as an adjunct to its EMC World conference.

Here is the thinking behind EMC’s evangelism of the number cruncher: The move to big consolidated data centers, and cloud computing, will speed the creation of ultra-large databases, holding information on everything from sensors in stores and infrastructure to what consumers are saying on Facebook and Twitter. “We have 1000 customers that have bought a petabyte of storage,” Burton says, “we think over the next decade that will grow to 100,000.”

via blogs.forbes.com

Here come the Data Scientists.

One of the many things going on at EMC World this year is that we're trying to get a massive gathering of industrial statisticians together for a Data Scientists Hootenanny.

45 days to go, then you get to see what for.

Storage Efficiency with “Awesome” FAST Cache

Have been testing various database workloads and virtual machine workloads on a CLARiiON Storage loaded with FAST Cache and must admit that the results are “AWESOME” !! To be honest we never thought Applications would be able to utilize so much amount of cache since we were using Systems with 8GB and 16GB Cache upto now . Everything has been proved wrong post the tests and here are some results that would blow you off .

Please find below some findings from Unisphere Analyzer on various application and VM loads.

via sudrsn.wordpress.com

I love customer proof points. 80% of the IOPS in this workload being served directly out of FAST Cache.

FAST Cache works and we've even more stuff on the way.

EMC's New Storage Line Poses a Threat to NetApp

Over the years, NetApp’s strength in desktop virtualization and its unified storage strategy under a single OS has helped the company boost its external hardware market share. The firm controls nearly 25% of the small-to-medium storage market and has been the dominant player in the cloud storage offerings and lower priced data storage products competing with firms like IBM and HP that also compete in this market. ^[1]

With EMC pushing into the SME market aggressively, NetApp may witness flat or even a decline in market share depending on the growth of this market and competition. We currently forecast NetApp’s share in total storage market will increase from just over 11% in 2010 to around 13% in the coming years. However if competition from EMC and others results in stable market share, this would imply about 12% downside to our price estimate for NetApp’s stock.

via seekingalpha.com

I read this earlier today and then read about NetApp fracturing the song and dance they've sold to Wall Street for years by buying Engenio, then read the claims on Dave Hitz blog that they're going to prevent overlap by not adding any features to the product line.

Of course NetApp's story has always been fractured when you looked beyond what they showed in the slideware. ONTAP 8 and ONTAP 8 Cluster Mode being two products sharing an installer, ByCast trying to clone Atmos Virtual Edition by shipping as a VM in front of FAS and now whatever they've acquired with Engenio.

As is usually the case poor Dave Hitz gets woken up and trotted out to claim all sorts of things any time NetApp gets brain fever. Previously they were "redefining IP litigation", by filing in patent troll city, now it's about spending $430M to buy something they say they won't do anything of note with or advance in any meaningful way so that shouldn't distract them from a business Wall Street thinks is starting to come under pressure.

Engenio will either wither on the vine due to being ignored, as is what happens with the majority of NetApp acquisitions, or, NetApp will have to split their development effort to advance the product line and make it worth buying.

There's more upside to EMC from NetApp buying this than not buying it at all. 

EMC has always done parallel development at a massive scale but over the past 15 months has started doing less of it as it develops products using storage software components.

The VNXe being entirely built from nothing but storage software components.

NetApp on the other hand are starting to defocus at a time when EMC is incredibly focused.

Game on.

Survey – Virtual Backup Servers | The NetWorker Blog

I have to admit, I have great personal reservations towards virtualising backup servers. There’s a simple, fundamental reason for this: the backup server should have as few dependencies as possible in an environment. Therefore to me it seems completely counter-intuitive to make the backup server dependent on an entire virtualisation layer existing before it can be used.

For this reason I also have some niggling concerns with running a backup server as a blade server.

Personally, at this point in time, I would never willingly advocate deploying a NetWorker server as a virtual machine (except in a lab situation) – even when running in director mode.

Let me qualify: I consider ‘director’ mode to be where the NetWorker server acts almost like a dedicated storage node – it only backs up its own index/bootstrap information; with all other backups in the datazone being sent to storage nodes. Hence, as much as possible, all it is doing is ‘directing’ the backups.

But I’m keen to understand your thoughts on the matter.

via nsrd.info

Preston poses an interesting question and I have my own thoughts on it which I'll hold onto until I see the results but needless to say EMC has NetWorker customers running NetWorker in a Virtual Machine and on Blade Servers today.

As per usual one persons sweet can be another persons posion.

Building Out Ahead’s Data Center Lab with Isilon

Back to the install process… Once we had unboxed everything and racked and stacked, we just needed to cable everything up. Beyond the traditional redundant power supplies, each node connected to the Infiniband switch via the included Mellanox cables, and I attached a single CAT6 cable for external connectivity. At a later time I’ll add another for high availability, but I was eager to get everything up and running!

We attached a serial cable to the first node of the cluster, fired up Putty, and we were off. We designated the bottom node as node one, and added nodes vertically from the bottom up. This ensures a consistent IP assignment scheme relative to physical placement.

Your cluster creation starts with the first node, and once you’ve defined basic parameters like cluster name, public & private IP addressing ranges, and DNS, the heavy lifting is *done*. I can’t stress how simple it was. IP address on a per-node basis is automatic, based on node number and the address ranges you’ve defined.

One note here – the cluster name appears to also be used when joining your cluster to an AD domain (for CIFS use), and I assigned one that was more than 11 characters long. Fortunately, I was able to go back through the GUI once we’d built the cluster and change it. Took but a few seconds to do.

From that point, all that was left was to add the individual nodes to the cluster, which was even easier than the cluster creation process. You have two options to accomplish this: you can attach a serial cable to each node and join the cluster, or you can do it entirely from the buttons on the front bezel of each node. Either way is dead simple; the nodes automatically detect the cluster, and you just need to confirm.

Granted, we had no data on our nodes, but each node took about 60 seconds to add to the cluster. We just kicked off the process and watched the web GUI update and include the new nodes in to the dashboard. Pretty amazing when you consider that is all it takes to linearly scale both capacity and performance.

via ThinkAheadIT 

Alex gets into some of the nuts and bolts of the Isilon out of the box experience but it's good to see him write what I discovered for myself.

It's insanely easy to deploy and scale an Isilon NAS Cluster.

And when you find that cool wait until you decide to remove or upgrade the hardware in a running cluster. Select the nodes to be ejected from the cluster, watch the data autobalance off those node(s) before you unplug the cables and pop it out of the rack.

And that's why Isilon is the leader in Scale Out NAS.

VNX+VG8: More than half a million CIFS IOPS

Following up on the record breaking NFS SPEC result (And the cat urine reaction from competitors who wrote yards of text about it all while stating “they didn’t care”), EMC has now posted it’s record breaking CIFS SPEC result.

661951 IOPS with a 0.81msec response time.

As Barry pointed out response time does matter.