dlewis23

We're testing your internet... not showing you fireworks... stunning graphics get in the way of the test more than anything you've described.. native or not, it's a waste of resources. Why? So it can look pretty. That's the furthest concern from my mind...

Except people don't like things to look like its 1999 anymore.

If you design it properly it doesn't get in the way it actually enhances the experience because you can show users things in a way that looks good but they can understand were just text suffocates the experience.

Figure out how to make a beautiful interface that doesn't get in the way (and you can actually do this) and people will use your app more. I've got a lot of hands on experience with this, in the latest update to one of my apps I improved the UI a lot, didn't change anything else and now I'm making 4 times as much on app sales and the app is getting used 8 times as much. All because I made the UI better.

For you, making a stunning looking website should be #2 on your list just after test accuracy. The UI is what the user sees and interacts with. If it looks like its from 1999 they will use your app/site like its 1999

If you're waiting it's because of the page loads in between things right now... I'll explain in the next section...

Easily solved with more development... development time that I think is better spent on the web application. ...work in some AJAX. Problem solved. ... It's not like its even a problem right now. Seriously, how long do you wait for your results page.... if it takes THAT long then you're getting to see a SECOND benchmark while you wait for something to load that shouldn't take more than a second. Even when I'm on the shittiest connections the results load instantly for me, so I really don't know what you're talking about there.

I'm not really talking about page loads i'm talking about when you start a test all you see is a progress bar. I have to wait for progress bar then the results page. That would be great for 2005 but its 2011 people want real time results. They want to see the results for the connection instantly from when they push start.

Since when do people hate websites on iPhone and Android? NO NO NO.. you've got it WAY wrong... people hate websites that don't view or or work correctly on those devices. I'm sorry but I'm personally in my browser 100X more than apps. I search the internet before the market or app store. --- There are obviously people out there that think differently from me but my question is... if websites worked better, would they still seek out the market or app store?

Since always. Apple tried to do the mobile app thing it didn't work out to well even for them. Mobile apps are slower, they can't do a 10th what a native app can do. And they are all very basic.

I don't care how much better you make mobile websites, they aren't going to be even remotely close to being able to do what a native app can do anytime soon.

There is a reason why I didn't build a mobile website for my sites and instead built an app.

Yeah, I remember that poorly executed genius idea that came too far ahead of its time... yeah I know what you're talking about. We are on a TOTALLY different infrastructure now. Once symmetrical fiber is commonplace ALL BETS ARE OFF. Video games, graphics editing... I don't care what it is. You'll still be able to compute locally but WHY? If you can do the same thing with nothing more than a input device and monitor. If you think I'm wrong... read this in 5 years.

Having a fast enough connection is only half the problem. The other half if the browser it self. Browsers are insanely slow at executing code. Thats why web apps are still fairly basic in the type of apps you can have in a web browser. They get faster at it all the time with every new version but they are a long way off from being even remotely close to how quickly native code can be executed.

That's true... but why not just build a really bad ass web application... then encapsulate it within a lightweight app that renders the browser? If exposure is the only real benefit...

Been there. Done that. Trust me it doesn't work.

It goes back to the fact that the browser is way too slow. A user will start the app and now has to wait. Because they have to wait for web-kit to initialize, then load the page your telling it to load and then if you have to click anything you have to again wait for the page to load.

It's not the way to do an app.

... I just want to make sure my time gets spent in the right areas. I don't want to have 100 different versions out there... I want ONE unified website. Not a website an two apps, which would surely yield different results from each other... that's not a benchmark.

So do what I did. Build an API. Then everyone is running of the same basic code base on the server and the only thing that is different is the front end on the App and website.

As the person who has the #1 iPad Speedtest app. By not having a dedicated app for TMN your missing out on a lot of faces your Logo could be in front of. As well as a lot of money.

well, not exactly but it will be less accurate because it won't be within real world circumstances so it would be a lot like doing a test from a site like speedtest.

That is actually incorrect. It will be more accurate to what your connection is actually doing. Since phones and tablets have really low power CPUs they are unable to max a connection above a specific point this number becomes even less when you add the browser into the factor since the browser engine is really slow.

Just for example the iPhone 3G was only able to get about 5.5 Mbps in the browser no matter how much faster your connection was. But if you ran a native app you could get to about 9 Mbps and thats what the phone was actually able to pull from the wifi connection. The slow CPU combined with the slower browser engine caused inaccurate results.

Since most things that use the internet connection of a phone or tablet are done outside of the browser, the result you get in the browser would be inaccurate to what you are really seeing in the real world.

I want to build iOS and Android apps for TestMy... but I still don't get it. Why use an app for something that can be done through the browser? Is it that hard to type "testmy.net" --- hell, in many cases you can just type "testmy"... why waste your local resources? Let my server run most of the program for you. ...am I missing something? Am I getting too old to get it?

Because it can be done in a native app better. With a native app you have access to more resources and more API's then you do in a browser. For you, having direct access to the devices network connection API not through the browser will allow for a more accurate speed test. Having the ability to skip the browser makes a big difference because now your not dependent on slow mobile device browser that really can't process much information compared to a desktop computer.

Also since you have direct access to these API's you can know what type of connection the user is on WiFI, 4G LTE, 3G, Edge ect. ect. and you can taylor how the speediest runs depending on the users connection.

The next benefit is your not limited to the browser and load times so you can do what ever you want with the UI and its not going to effect the speed test results. You can build a stunning UI with as many animations and types of animations you want. Show results in real time with out effecting the speedtest result.

The way you have the test setup now I have to wait for my results. This isn't a doctors office, why do I have to wait for my results? Build a native app and I don't have to wait. I can see them in real time.

You also have page refreshes right now, several of them if you do and upload & download test. People hate waiting for a page to load this is why HTML5/Javascript apps on the desktop are so popular no more waiting for the page to refresh.

With native app no more refreshing, you can do a download and upload test every time with a seamless transition from the download test to the upload test.

I just hope that the world realizes one day that you don't need apps... AT ALL... FOR ANYTHING. (except offline) That goes for personal computers as well. Everything can be done through the browser, especially more so in the future.

Except that day is not today or tomorrow or next week or even next year. Running something through the browser is a least 10 times slower then running native code. Also the browser does not have access to even 5% of the API's that are available to native apps.

Remember WebOS? They tried to run everything through the browser and failed. Mostly because the browser is just too slow for running apps and you really can't do a lot of the stuff you can do with native code.

Do people like having a bunch of crap they need to update all the time?

Why waste your disk space?

Well IDK about Android but in iOS5 everything gets updated automatically in the background. And even so updates have never been a problem for people since the device tells you an update is available and there is a update all button that people just tend to push and let the updates happen on there own.

And as for disk space it really doesn't waste much of any disk space. My latest iPad app with a ton of high quality images is only 9.7 MB. If I really wanted to I could get that down to about 6.5 - 7 MB but I don't want to loose any bit of quality in the images because I can see the difference.

Remember this, people hate websites on iPhones and Android devices they are more likely to go to the App Store and find something they want then browser the web for something on there phone.

If you built a good speedtest app you would have 100,000 new users in a month.

Thanks for everything Steve, you've made a big difference in my life

Wether you like Apple or not. He made a big difference in everyones life.

D why would you use eAccelerator over something like memcache , is it more image friendly ?

In my case it just seems more friendly memory wise, a little faster and works best with one of my PHP scripts. Ever file on my site is downloaded through a PHP script and memcache just seemed to mess that up. It was caching the script for files that were for a different file. eAccelerator doesn't do that.

Please explain what in the hell you guys are talking about... lol

Ahh, you'd think I'd know what you're talking about... but I've never heard of those modules.

You would think of all people on this site at least you would know what the hell we are talking about. lol.

interesting i might have to try that one out, does it work well with worker? or only prefork?

Should work just fine with worker but I always use perfork just seems to be more stable for what I need.

Neither. I use eAccelerator.

I can't see why you couldn't create the partitons on a RAID 1 but If something were to happen I don't know if it would maybe cause a problem when it comes to restoring data.

Every time I have done any kind of RAID Ive always kept it as one and had the OS/C Drive on a separate drive.

Can't hurt to try. Just create the RAID 1 first then and partition it using Disk Manager.

Yea it is nice and clean, plenty of room for heat removal.

There really isn't much heat to remove out of this thing. The CPU is running at a max of 28C when I had the lid off. And the drives never go above 32C. It will probably run cooler once its in the DC since they keep it at 75F.

i was just thinking the same thing myself CA3LE, it looks like a very well put together machine there lewis not bad cable management either although i'm sure that wasn't a main concern but it looks very good

~ TriRan

Cable management was the primary concern of the system. Since the server is going to be far away I won't have quick access to it, so I needed everything to be cleanly laid out and labeled so that if something did happen like a drive failure I can tell my host what drive specifically to remove and its exact location.

I labled every cable at both ends saying what drive goes were and I labeled the drives with there serial number and disk location in the server, from Parity to disk 8.

This setup is pretty fast too. WIth out doing any customization on it yet, I'm getting about 50 MB/s read and write to it and thats through a real cheap. as in free Netgear router. Should be a little quicker once I do some tweaking then a lot quicker once I get on a real switch.

What is the pricing for it all? (The Datacenter stuff, hosting your server there and whatnot.)

Thanks,

EBrown

The price of the server with the software was about $900. The drives were the most expensive thing. As for the monthly hosting cost I can't give that cost away. I never say what my operating costs are.

I will get most of the cost of the server back. Since this will be replacing my current storage Items that I will sell once everything is on the new server.

I finished the build this morning after picking up a few SATA cables and making some short brackets for the SATA Raid Cards.

Overall not bad, 16TB of usable protected space. Will be working on #2 real soon, probably after this ships and is all setup in the datacenter. Now I gotta build a damn box to ship this sucker.

Here are some pictures:

Sure , tried installing several times that way, it always fails 40 minutes later complaining about not being able to download items for the install.

I was not aware the USB stick ( if you can call it that ) , was for emergency repairs, and what recovery partition are you speaking about ? I take it Lion has implemented something from microsoft now ?

Yea your only suppose to get the USB Stick if you have a problem were the recovery partition doesn't work.

The recovery partition is everything that is on the USB Stick on your Hard Drive. It's not just recovery, its everything maintenance wise that use to be on the DVD, disk utility, terminal, password reset etc. etc.

Burn the DVD at a slower rate like 2X and do not verify after. Don't ask why, but thats what fixed the problem with the DVD for me and many others.

Anyhow , I was going to get the USB stick to do a clean install until i read on apple site that the damn thing isnt even the OS , it's a key to download the OS , and snow leopard is a prerequisite to lion. wtf apple ? Geezus , no wonder Steve left !

You know you can make your own DVD to do a clean install right? You don't even have to install it first, just download. The USB stick is only really for people who have majorly screwed up there computer to the point where the recovery partition won't work.

If what you say is true , that rossetta has been removed , then a quick search will reveal there were quite a few that still used PPC apps. Yet it also means I'm safe , I'm just interested in progs such as transmit , coda , PS5 , navicat , paralells ect ,nothing out of the ordinary or too old.

What is all this about slowing the machine down significantly ?

Hey if people are still using PPC apps and upgrade to lion thats there problem.

I use a lot of the programs you also use, and have no problems under lion.

As for slowing the machine down I haven't seen that or really read that on any kind of large scale. The only 2 problems I still see are with Webkit2, if I don't restart Safari every few days it starts to really suck up some ram and you see a performance hit within the browser. The other problem is with the Ctrl Zoom, sometimes it just doesn't want to work.

So is Lion gotten any better with the release of 10.7.1 ? Or is it still borking the brains out of everyone ?

Iv'e read MS Office does not run on it, as well as Photoshop and about a thousand other apps, was this user error as usual or are all my apps useless if I update to lion ?

I have yet to see any apps that will not run on Lion.

When you hear someone saying office or Photoshop won't work its because they are using a really old version of either Photoshop or Office. Rosetta was removed from Lion so old PPC apps don't work. But who still uses PPC apps? you would be using versions from 2005.

damn

What, Never seen that many drives at once have we?

Oh, ok. That makes more sense.

So what features does the RAID the OS makes have? Parity, etc.

Thanks,

EBrown

Parity, really easy recovery of a drive does fail. Boots from a flash drive which is awesome. It has active directory support, cache drive for better performance, can handle up to 21 drives on a single setup.

I picked up the drives today. Got a good deal at Compusa on 2TB hitachi Drives.

So thats what 9 drives looks like on my table. lol..

Alright, that sounds good.

Which RAID mode are you using then? (Obviously not 0, 1, or 2 by the looks of it.)

Thanks,

EBrown

None. I'm using UnRAID. The OS UnRAID makes its own raid at the OS level, its not like a traditional raid. It works just like a Drobo you just supply your own hardware.

Ah, that makes sense.

So is speed really an issue for it then? (I only ask because you may be able to find a cheaper HDD of the same size, just slower.) Otherwise, I think the Seagate 2T's you picked out will work fine.

The only issue I can see right now, is the lack of enough SATA power cables on the PSU. (Unless you have Sata to Molex, in which case ignore this.)

Here is a PSU with 9 SATA Power Cables out of the box: http://www.newegg.com/Product/Product.aspx?Item=N82E16817207013

It's slightly more expensive than your first choice though, and again, if you have Sata to Molex, yours will work just fine.

Thanks,

EBrown

Speed will be an issue, but it will just be limited to the 1 Gig port. The goal is to get about 70 Mb/s (560 Mbps) read/write, which on this type of setup with parity.

I take it this OS is ready to accept scsi jbod as well ?

Yes, but it will turn them into a Raid in one big drive on its own, since thats what UnRAID does.

I recommend 2T's, and I recommend Western Digital. (I have had huge issues with Seagate, mostly with them failing after 3-12 months, almost ALWAYS. I have had the same WD 80G for 7 years, and it still works 100%.)

Also, are your RAID cards SATA-II, or SATA-III?

Thanks,

EBrown

The problem with going with WD drives is there green drives are 7200 RPM and they use more power then the Seagate ones. I have been running 4 of the Seagate green drives in the Netgear NAS I have for a while now with no issues. It seems as long as you keep them cool they are really solid.

The motherboard is SATA-3 the raid cards I'm getting are SATA-2. The mix doesn't really matter much since the entire setup will be limited by the network port speed of 1 Gbps.

looks awesome although if it were me i'd be going for a tad beefier CPU like the http://www.newegg.com/Product/Product.aspx?Item=N82E16819103903

it has a lot nicer L1 and L2 cache for not a whole lot more, i actually own one of these chips and it's in my brothers computer i used to use it as my main CPU back a year or so ago

For this type of server that CPU is beyond overkill. The CPU I picked out is overkill. All this server will do is store files, it does nothing else. The OS it runs (UnRAID) is an extremely lightweight Linux that will never tax the CPU. I've seen 24 drive media server builds use a Celeron 430 and they never run into any kind of CPU issues.

Plus I have to keep power usage to a minimum. Since I will be collocating this server I have to pay for power usage.

EDIT: Forgot to add, the Sempron 145 Processor can actually be unlocked and turned into a Athlon II x2 by just changing a bios setting. I won't do it because I need to keep heat and power usage down, but the option is there.

As some of you know I run a file/image hosting service and my current storage solution is quickly being out paced by the sites growth. For the amount of files I'm hosting the way I currently store files across many drives and NAS devices is very inefficient. So I have to build something of my own. This will be the first of 2 storage servers I have to build over the next month.

The Hardware:

Case: http://www.newegg.com/Product/Product.aspx?Item=N82E16811147156

Processor: http://www.newegg.com/Product/Product.aspx?Item=N82E16819103888

Motherboard: http://www.newegg.com/Product/Product.aspx?Item=N82E16813157199

PSU: http://www.newegg.com/Product/Product.aspx?Item=N82E16817371030

OS: UnRAID

Hard Drives: 9x http://www.newegg.com/Product/Product.aspx?Item=N82E16822148681 or http://www.newegg.com/Product/Product.aspx?Item=N82E16822145493

Ram is just going to be a basic 1GB stick and I will have 2 PIC-E Raid Controllers so I can have all 9 drives. Depending on what kind of bulk deal I can get will decided what drives I end up getting, should find out tomorrow or Tuesday.

In the end if I go with 2 TB drives I will end up with 14.4 TB usable space, using 3 TB drives I get 21.6 TB usable space but at almost double the price. I'm leaning toward 2 TB only because the cost savings can then go towards the second server.

I will have all the parts this week so this should be a fun project overall. Pics of the parts, build, and finished product coming soon.

How big is the SSD you went with? And did you do the easy route and just replace the main drive with it or the hard way with using the mount under the optical drive?

You all need to quit your belly aching and admit that Linux is the clear winner.

This coming from the guy who bought 2 Macs in the last year or so.

The thing is, you're supposed to be more mature, yet you keep coming back.

Fuck you and your direct insults. I don't need that shit, I'll go to my High School if I want it.

And thanks CA3LE, for unhiding that post. Mudman had calmed me down until you unhid it.

Goodbye Testmy.Net,

Elliott Brown

Hey your back! I thought you said you were quitting? Or is this going to be like when a smoker says he quit, yet he's lighting up right in front of your face.

Yea I'm suppose to be more mature yet your the one saying "Fuck you" to me, so I should just let you get a pass because your immature? Thats not how the world works. You have got a lot to learn.