@Dashrender said in Scale Computing Brings First Fully Featured Sub-$25,000 Flash Solution to SMB Market:

@PSX_Defector said in Scale Computing Brings First Fully Featured Sub-$25,000 Flash Solution to SMB Market:

SATA is perfectly fine for 90% of what people do. It's the 8% that need something more that would need SAS based while the last 2% will need PCI-E performance.

With numbers like those, ML seems like an odd place to be talking/worrying about it. Also, are the last 10% really looking at a Scale Cluster? I suppose some percentage of them might be.

90% of what people do, not 90% of people. It's a much higher percentage of people. That's why the Scale HC3 tiering system is such a good fit, we believe. It allows the majority of your storage to be tuned to sit on the SATA drives, which are perfectly fast enough for 90% of your needs, and lets the 10% of your needs that need to be on high performance SSD to sit there without needing two different solutions.

And with our heat mapping technology we help to tune the workloads for what is used rather than forcing you to pick manually for all workloads. You can override this with manual priorities, but on its own it self tunes.

So our hope is that the 90/10 split which is a good way to think of it actually makes Scale ideal for the majority of users because they have the 90/10 mix rather than in spite of it.

@Dashrender said:

Most storage devices in this range also lack the support options that enterprise servers do.

This sentence is the third italicized block of text seems odd.

Fixed, thanks.

@tonyshowoff Now that's just painful to think about.

thanks, fixed.

@scottalanmiller Thanks for this post and answering all my questions. @travisdh1 Thanks for answering all my questions as well. Good Thread!

@Jason said:

Almost any NAS can do iSCSI a SAN is simplier than a NAS..

Exactly, but a few scale outs don't because of the scale out aspect.

@wirestyle22 said:

@coliver said:

@wirestyle22 said:

I have to look up SSL certification on CentOS 7 now. I've only ever done it on our Barracuda SSL VPN which was just an upload. Here we go

I think you need to look up SSL for Apache if you are trying to encrypt ownCloud. I think someone has a guide on here somewhere about it.

I think @JaredBusch has it in his 8.2 guide. I think he did self signed though. I'm unsure of how to do this on linux with digicert or something similar

It was a self signed for the example. but I use letsencrypt myself. Some clients using ownCloud have StartSSL certs.

@scottalanmiller said:

@Dashrender said:

What's your take on this java utility?

If I recall previous discussions, you're completely fine with server side Java things, it's client side that are horrible.

Considering this is an admin tool, is this good or bad that it requires Java?

Admin Tool = Client Side

So for me, I think it's a bit of crap there. As an admin tool I'd find it one step better than if it was for "generic" users. But in this role, it is still for an end user, just a power user (me.) And I think that any use of client side Java is rather bad. Not the end of the world, but not good. This isn't a new utility, so we can excuse the choice more than we could with a new one. But I'm certainly not planning to deploy it. There has to be better options.

I feel exactly the same way. Was mainly wondering if you did too - and well, you do 😉

http://www.theregister.co.uk/2016/02/26/canonical_in_zfsonlinux_gpl_violation_spat/

I forgot about this topic and found it mentioned in a conversation. This thread was a great resource that never got linked anywhere useful. Now to figure out how to make it more referenceable.

There are systems that do this with version control. You could do this with GIT or Mercurial, in theory. But eventually a human always has to sort things out.

@scottalanmiller

@scottalanmiller said:

@SeanExablox thanks for all of the info!

anytime!

@ReneeHayter welcome to the community!

@Breffni-Potter said:

Companies like Dell/HP/Lenovo are getting around the SSD issue by adding 200% markups to them 🙂

Don't get me started!!!

http://community.spiceworks.com/topic/900545-purchasing-hardware-from-dell-a-well-intentioned-rant?source=year-in-review

(I was mad, which is rare... I think I had just gotten off the phone with my sales rep lol)

I see the use case for these in databases and almost solely there (until they get cheap.) We would need to know a lot about their performance, longevity, reliability, durability, cost, etc. before we could speculate much. Seven times faster is a big deal, that means that things like FusionIO are going to have to rethink what they are doing. And the SATA and SAS bottlenecks would be useless and we'd be forced onto direct PCIe or faster for any hope of getting the speed out of these things. But changes there have been coming regardless because already are drives are bottlenecked by the connection technologies.

If these end up the same price as current flash SSDs but are seven times faster and all other things are equal, that's amazing and we will switch right over and the flash drive is dead. But I suspect that there are many factors that we have yet to discover. And, right now, it's just a long off dream. We will have to see if they really bring these to market.

By and large, though, systems like file servers and application servers are no longer bottlenecked at the datastore layer but on the network, so I would not see most workloads having a reason to upgrade, yet.

@craig.theriac Ok, hats off. That is cute. And some great marketing!