Maxime Rousseau

Paypal is a blessing, because it allows anybody to send or receive money online, at no cost, and without the need of having a credit card. However, the painful part is seeing your money take DAYS to make the trip from your bank account to your Paypal recipient. Us Canucks have it extra hard:  American bank account transfers take only 3 to 4 business days, which can easily translate to a full week if you don’t start your transfer on Monday, and the delay is 6-8 business days for Canadian bank accounts. Imagine a situation where I’m buying something online from some eBay store in the US. 10 days for Paypal to clear if you count a weekend in, and then at LEAST 4-5 days for shipping and processing and whatever else the stuff has to go through before getting slammed on my doorstep by an idiot dressed in brown. In my book, that’s 15 days before I get my stuff. 15 days is over 2 weeks, and two weeks is half a month. Half a month is a 24th of a year, and a 24th of a year is pretty darned long.

Of course, one can just add a credit card to his/her Paypal and benefit from instant payments, may they be from credit card or from your bank account. However, not everyone on Paypal is of legal age to own a credit card, and not everybody has a mom or dad willing to supply the plastic. Up to this day, we poor young customers of the internet were cursed with painful waiting times, and rejection from many e-stores and sellers who don’t like the concept of waiting for a payment to clear. Until now!

This time of waiting is over. There is a way to “trick” Paypal into fronting you the cash for your online purchase while all the processing is going on between Paypal and whatever financial institution you deal with. The basis of this little bypass is a new type of card released by Mastercard, the Vanilla. This card, which I see little use if you ignore this little hack, is basically a universal gift card. It carries a predetermined monetary value, however unlike your regular gift card, it’ll work wherever theres is a credit card terminal, which includes real brick and mortar stores, but also online stores. What’s special about it is that it has a card number, a security number, and an expiration date, and it’s treated like any other credit card. From there, I’m pretty sure that you can see where I’m going with this.

To activate immediate eCheck payments on your Paypal, all you have to do is get one of those nifty little cards at any good corner store (I got mine from Couche-Tard, the Canadian equivalent of Circle K, so they should stock them), and bind it to your Paypal account like you would with any other credit card. Paypal eats it right up, and within minutes you’re sending money in a matter of seconds, anywhere in the world. Amazed, you’ll most probably start bidding on hundreds of eBay items like I did.

Be warned though. How the system works is that your bank account is still considered the primary source of funding, and your credit card is added as a backup source, if the bank account transfer bounces back. Since your “credit card” has a prepaid limit, this means that if you “acidentally” spend for more than you have in your bank account, then both your prepay and bank account will bounce back. You’ll be fucked. And by fucked, I mean having your Paypal account suspended.

So be careful.

Maxime Rousseau All, How-to, Internet

Most people who buy pre-built computers usually find comfort in XP’s user experience, and are ready to sacrifice some RAM to get that comfort. But as an enthusiast, it brings tears to my eyes. Why would anybody want to waste RAM? You payed for 4 gigs, why are you only using 3? The solution to this problems is obviously getting people to adopt and keep 64 bit operating systems, and there isn’t a million operating system vendors out there; the one company that can make things change is Microsoft. What MS doesn’t seem to realize, is that they have the tool to finalize that transition out in the hands of the public right now: Windows 7.

Read the complete entry on my OCN blog.

Maxime Rousseau All, Hardware

I was at work the other day, a lady came up wanting a preventive checkup (included in her extended warranty) and AV reinstall on her Mac. As funny as I thought the idea of having an antivirus product on a Mac was, I sold the lady the latest Norton AV for Mac, and took her computer in for a parts check. Turns out PC-Check doesn’t have the right drivers to run on a Mac, as the keyboard stops working once it boots (tested with 6.05 and 6.21), but a custom launcher with command line arguments works flawlessly.

So yeah, in the end, all the parts were OK, I did some maintenance (which on a Mac can be resumed to updates, blowing some air in the cooling slots), and ran into some problems when re-activating the Norton product. Our tech-bench version of the software doesn’t have any activation keys printed on them! For those of you not familiar with Norton products, the product key stickers look like this:

In the two boxes that I have unwrapped, both of the product booklets came without the yellow sticker containing the key. Concerned, I fired up a livechat session with a Symantec rep. Here’s the integral transcript for your very own reading pleasure, I’ll  be highlighting the important stuff.

As soon as I got a hold of somebody, I inquired on how I could find the customer’s product’s activation key elsewhere than it usually is. I had already tried entering other numbers I saw on the box as the activation key, with no avail. The only question that I got was for a billing address which the dude said was for quality assurance purposes only, and that I sadly couldn’t supply as to protect my customer’s identity. The dude’s response?

PrasadDuvvi(Sat May 23 12:42:42 EDT 2009)>
No problem. 

PrasadDuvvi(Sat May 23 12:43:37 EDT 2009)>
The activation code is (*********)

PrasadDuvvi(Sat May 23 12:44:38 EDT 2009)>
Please try to activate the program and let me know

It was that easy to get a key. I was totally puzzled. I actually asked the guy how come he gave me a key without asking for a proof of purchasing and without questionning me any more than he did. His response was that he simply carried out what I wanted and gave me a key. 

What is to be learned from this? That the OS-X antivirus products are so unprofitable that you can give them away for free and not see a difference. I understand that they guy probably only wanted to get rid of a quick case by giving away something that wasn’t worth very much, but what kind of message this give to consumers like me?

Now this brings me to think that maybe I could do this with other Norton products. I haven’t tested it yet, but my main server OS disk is in dire need of changing and I would really like a free copy of Ghost. I’ll keep the post updated if there is any avail. Post in the comments if you try it out!

Maxime Rousseau All, Internet, News, Software

I love Google. Ever since I’ve registered a Gmail account, I’ve given more of less every one of their services a shot, from the blog search engine to the now defunct Google Browser Sync. Overall, I have found their services reliable, practical, and just kick ass in general. 

I have already been using IMAP Gmail on my iPhone, and like every proud iPhone owner I never miss a chance at anwsering/correcting people by whipping out Google from my pocket. I thought I was in the geek-nirvana… always connected to whatever I wanted to be. Right up until I stumbled upon Google Mobile Sync. 

The concept is pretty simple: one phone, one Exchange server, and all your contacts and calendar entries are syncronized from whatever your smartphone is to the great big internetz. For the more visually inclined, have a video.

GOOGLE HAS FREED US OF CABLES, THE ARCHAIC MEDIUMS THAT USED TO UPDATE OUR ADDRESS BOOKS! Imagine you’re on the job, Mr. X sends you a very important business email telling you that he wants you to call back. No problem, you add the dude as a contact in your GMail account (because you are using your business POP3 on your GMail for superior organization of course). Then, two hours later on the train to whatever important place, you want to call up the dude, and his number has MAGICALLY been pushed to your address book. Badabing badaboom. No more messing around with awkward browsing in your Gmail contacts via 3G. You are free of wires, or typing in long names like Adolph Blaine Charles David Earl Jr the Second or numerous phone numbers on your iPhone’s teeny tiny on-screen keyboard which makes typing a 9 instead of a 0 so easy. I think that’s pretty cool.

Or consider this: you’re a busy man, the manager of whatever facilities, let’s make it a hospital. Like any self-respecting busy man, you have a suave secretary who juggles with managing your schedules, returning your phone calls, and typing up your stuff. With Google calendar, you can share a calendar with Nancy, the said secretary, and have her manage all of your time, remotely, while not having to place a single phone call and being updated within literally SECONDS (I’ve tested this… push is that fast). Meeting with CEO of SOandSO Corp over, where to? Whips out iPhone. The press conference at the Imaginary Hotel isn’t on my calendar anymore… it’s been cancel. But hey, I’ve have a flight to Toronto to catch in an hour so I can help another hospital with their problems relative to whatever. Oh my, even the details relative to my flight are mentioned on my calendar. The obvious downside is that you don’t get to hear Nancy’s sensual voice every hour, but hey, it’s much better than to take quick, half-assed notes in a paper agenda, isn’t it?

I know, MS Exchange isn’t anything new. But having your very own little private slice of one of these for free certainly is. So for the love of god, just get your sync on. It takes about 5 minutes to configure, and the most difficult thing for me was to clear out my Gmail contacts.

I love centralizing. Long live Google.

Maxime Rousseau All

It’s been a while since I’ve posted here. In between school, shitty homework and working 30 hours a week, the little but of time that I used to devote to updating my blogs is now down to very little to none. It seems that it’s when there’s the most stuff going on around you that’s post worthy that you can never find time to post. Anyways… 

Energy drinks. It’s the new thing, everybody’s doing it. Caffeine has been transformed in the minds of the general public from a jitter-inducing chemical in coffee and tea to a full fledged drug popular amongst the younger people. Everybody under 30 has consumed an energy drink at least once, and according to what I’m reading, it seems that a lot of people depend on these beverages to keep their busy lives going.  Schools have started banning the stuff because it supposedly makes children a tad too excited. Give it a couple of years (months?), and it’s sur to become the next big social issue, right up there with the legalization of pot.

Like every youngster out there, I have been consuming energy drinks for quite a while. I got introduced to the stuff by a couple of cadet buddies who took it at biathlon competitions hoping it would improve their ski times, being handed over whatever was left in the piles over piles of cans of monsters that they brought with them and didn’t want to bring back home. I liked the taste, but I didn’t find the experience particularly enlightening, and I sure as hell didn’t grow wings. Eventually, once I started working some night shifts at McDonalds, I started buying my own energy drinks, and started experimenting with different brands and formats. At first it was the tried and true Redbulls and Rockstars, then Hype (which remains my favorite up to now), and by the summer of  ’08, I’ve had my fair share of those drinks, yet rarely took them out of work. 

But last year, I started taking more and more of the damned beverages. By winter 07-08 and up to October ‘08 , I was consuming caffeine on what some of my friends called an “alarmingly regular basis” : at least 4 times a week, and on occasions two times a day, when I really needed a kick. Being a caffeine sensitive person, I didn’t get much of an instantaneous boost out of drinking that stuff, however I would feel incredible tired if I spent a couple of days without drinking one. Buying 4 and 6-packs to save some money wasn’t a waste anymore, because  normally those wouldn’t even last a week. Between classes, at lunch, I would drink them whenever I felt slightly snoozy. Since then, I have stopped drinking energy drinks so much, but I still take on on occasions.  

The expansion of the market for these beverages in the last years has been enormous, new drinks, flavors, brands and variants popping up everywhere on a daily basis. But with every trend comes the anti-trend: think baggy pants vs girl jeans on guys. Now that the energy drink market is going strong, a company is starting to market it’s total anti-thesis: a relaxation drink. 

Called Drank, which is a pretty obviously a nod of the head at DJ Screw’s and later Lil Wayne’s infamous codeine syrup based Purple Drank, the drink promises to put you in a state of relaxation they qualify as “extreme”. The beverage is said to be entirely natural and made out of Melatonin, Valerian Root and Rose Hip extracts, although I’m not sure how extracts of a substance which the body secretes naturally and that fucks up your body’s internal clock qualifies as natural. Oh, and it’s purple and it tastes like grapes. 

While the concept of the drink is nothing extraordinary and perfectly acceptable, because it basically is just an over the counter, self administered mild sedative that can work wonders for busy people who ride planes all over the place all the time and need a cure for jet lag. However, the existence of this product in tandem with energy drinks just feels weird. The way I see this, some people somewhere in the States were having too much energy drinks, and one day decided to take action and invent something that would get them off their Redbull high, and BING, Drank! Are we just substituting an energy for a relaxation one? Or are we just making manipulating one’s energy easier, one slim can at a time? 

I can really see people abusing from this kind of availability of both stimulants AND sedatives. Take this guy for example. If you can read French, please do read the article, it’s pretty hilarious. To sum it up for those of you who only speak/read English, basically the guy did 80 hours weeks all the time, all while smoking a lot, drinking up to 8 c0ffees per day, eating crap all the time,  and downing a couple of energy drinks to wash it down. After a year of doing all of this, the dude dropped (almost) dead and had a cardiac arrest, at the sweet age of 32. Unfortunately, modern medicine was able to trick natural selection into letting this sucker live, and he is now on a crusade to end the reign of energy drinks over young people. He goes as far as to say that these drinks “kill people” and swears that he will never drink another super-caffeinated beverage again. 

This guy obviously didn’t need Drank to comfort him into drinking liters on liters of Monster, but how many people do you think will get caught in the trap where they’ll think they can down as much energy drinks as they want because Drank can help them get the sleep they need? What if all these people wake up one day with whatever part of their brains that secretes melatonin totally under-worked from drinking all that Drank, and can’t fall asleep without anymore? After all, isn’t it proven that sleeping pills are extremely habit-forming?

Don’t get me wrong, I’m not of those who oppose such drinks, but I just find it wierd that people have to rely on those products. We are but recreating less potent versions of cocaine (and other stimulant drugs) and opiates, how long will it be until people develop a resistance and want stronger  stuff over the counter? After tiny Hype shots, how long before people starting free-basing caffeine?

Maxime Rousseau All, Other

About may of last year, I built a server for myself, in hopes of centralizing all the storage drives of my house to one constantly on box and simplifying access to files across the four computers on my network. And sure enough, a couple of hours and hundred dollars later I had myself a kick-arse server rig which housed a RAID1 array for my personal use as well as 3 other random storage drives on which the members of my family could store their music, videos and pictures.

While having my music on a network shares was convient because I could access it from any machine on my LAN without running server software on my main box, I quickly found a flaw in my setup just a couple of months after buying it. Sure, all my music files where save on my server, but the iTunes library XML files were still stored on my local disk, which made re-adding all my music to iTunes after an OS reinstallation unbeleivable long (reading iD3’s on 50+ gigs of music isn’t what I would call quick), and reading the music from another installation of iTunes on another computer, say a home theatre PC, impossible.

While iTunes allows the user to choose it’s “iTunes Music” folder, it does not supply such an option for the library files. Ever since, I’ve been looking for a way to get my single library to work with many iTunes installations. Sure, there are some how-tos out there, but even the most popular solutions aren’t elegant and require third party software. Wouldn’t it be great if Windows, like linux, would allow us to bind symbolic links from one folder to another? Oh wait… there is such a function on Windows Vista. mklink!

This somewhat obscure command makes everything easier. With mklink, you can simply create a “dummy” iTunes folder in your My Documents folders, and have that folder to point to whatever the heck you want, the latter including network drives. Here’s how it’s done:

• First, you need to create a folder on your remote server that will be taking the place of the iTunes folder where all the XML files and album art is stored. Under XP, this is usually C:\Documents and Settings\Username\My Documents\My Music\iTunes and under Vista it would look something like C:\Users\Username\Music\iTunes. When that’s done, transfer what’s in your “iTunes” folder to whatever remote path you are using. In my case, my music files (the “iTunes Music” folder) was R:\Music\iTunes\iTunes Music, so as to not mess around too much with the directory structure I decided to create a folder at R:\Music\sync which would act as the target folder.
• Next up, you need to transfer all the stuff in your iTunes folder to the target folder, because you will need to delete the actual iTunes folder in order to create the dummy folder. Once the transfer is done, delete the iTunes folder.
• Lastly, create the dummy folder from the command line using the mklink command. The syntax is pretty simple for this command:
  

  mklink C:\dummy\folder  D:\target\folder

  So in my case, the actual command I used was this:

  mklink C:\Users\Administrator\Music\iTunes R:\Music\sync

  There are many types of links that you can create via this command, but for our us a symbolic link is enough.

Once that was done, my C:\Users\Administrator\Music\iTunes folder effectively pointed to R:\Music\sync as intended. iTunes blindly took what XML files were in the destination folder, and loaded my library on first try.

From then on, syncing other instances of iTunes with the current library is a walk in the park: make symbolic links from the iTunes folders of other machines to the same mounted network share, and voila! The other installations of iTunes will read into the library files thinking it is there own, giving you instant access to your music as long as the “iTunes Music” folder path is the same on all the machines. Changes to the library will also be written instantaneously on the unified files, making the changes available on all the other machines.

Before you start unifying your libraries however, there are additional things to take into consideration. Firstly, I would pre-configure the installations of iTunes as I do not know if certain aspects of the configuration (iTunes’s Music folder, adding songs to iTunes’s Music folder when added to library) are set in the registry or in the XML files. Also, keep in mind that you can not open the same file for read/write operations two times simultaneously, which means that opening two iTunes on the same library at the same time will most probably brick your library files if any changes are made.

Despite it’s limitations, this method is the simplest way I found of getting many instances of iTunes to share a same library. No waiting for dropbox or other automated backup software to do it’s thing, no maintaining of two sets of library files, just easy access to your music from anywhere on your lan. Obviously, the biggest bummer about mklink is that it only runs under Vista… but hey, if you don’t have Vista yet, you should really try it out, and if you don’t feel like trying it out, then the very interesting Windows 7 which is right out of bend is your next alternative. Sorry XP users!

Enjoy!

Maxime Rousseau All, Software installations, itunes, lan, libraries, library, management, many, network, Software, unify

Everybody knows that we, members of OCN, have a serious problem. Look at it from an average Joe perspective: we spend thousands on high end computers, and at every single occasion we go out and browse Newegg on a quest to find an upgrade for our systems, which according to us are quickly becoming obsolete. We then proceed to sell our memory, processors, graphic cards and all our other barely used gear, things which are sometimes only a couple of weeks old, in order to purchase the latest and greatest. While the common mortals still call a Pentium 4 HT/P965/7900GX2 system a top gaming machine, we constantly crave for more, more performance, more 3dMarks, smaller SuperPI times, and bigger e-peens. And after spending 50$ on Crysis:Warhead in order to be able to run the benchmark tool, we sit our asses on the chairs and play CS:S at framerates exceeding by 6 times our monitor’s refresh rates. We are hardware junkies, hung on PCB like a druggy is on PCP. 

Probably my best piece of to now. I spill my guts on what I think of the new Intel platform that features the new Core i7 processors and x58 motherboards. Read the entire article with feedback from the Overclock.net community here.

Maxime Rousseau Hardware core, epeen, Hardware, i7, intel, motherboard, ocn, overclock, overclock.net, platform, processor, repost

Every since I’ve joined twitter back almost two years ago in March of 07, I’ve been tweeting on a regular basis. I love Twitter not only because it’s a great place to dump all the thoughts and reasonings that are too short or stupid to blog about, but also because when one logs what he/she is doing on Twitter, there is no telling what follower will reply, and what kind of conversation will ensue. In that sense, I think of Twitter not only as microblogging, but also as a great big open SMS conversation. Thanks to my followers, I’ve found tons of new sites, and equal amount of funny/interesting videos, and shared tons with lots of interesting people.

However, Twitter has it’s faults.  If I had the power to change 5 things the service, changes would look something like this. 

1. Stable servers. The summer of 2008 was pretty rough for twitter, most likely because of a boom in member count and the increasing number of third party services that started to use the Twitter API (I’m thinking Twitpic, amongst others), the Twitter servers went through some serious problems. From what I understand, the app had some major scaling issues around that time, but I can’t quite understand if it’s a hardware thing or a Ruby on Rails thing. EIther way, the twitter server are much more reliable now than they were last summer, but I still occaisionally see the dreaded failwhale when searching or browsing some pages or using more advanced features. 
2. Groups. What’s a social networking site without groups? What happens when somebody wants to follow a hundred or more people, and keep track of all the tweets in a organized way? Group pages would allow me to see just the tweets of the group members, while would simplifying mass following. Similar to the @reply system, users could post an update to a group by typing *group before their tweet to post to that specific group. 
3. Proper implementation of the tagging system. The #topic system is widely used, but it seems that putting in the #topic tags has no effect on search results. If I search for “cats”, I get all entries containing “cats” and “#cats”, yeilding irrevelant search results in some scenarios. #topic tags also shouldn’t be a part of the already very limited 140 character limits. Tags should be splitted in two groups: #tags which are visible, which can be used in the following situations:
   

   Check out my new #bmx wheelset!

   And ##tags, which would be invisible in the tweet’s text, but that would still tag the post with ##whatever. #tags and ##tags could also link to a page where all recent tweets tagged with the concerned text, and could also include a neat little tag cloud.

4. A built-in video or photo/video viewer. Twitpic is becoming increasingly popular since the advent of the iPhone, and I can’t recall the number of Youtube videos that have been linked to me from tweets, so the next logical step is to add some sort of AJAX picture and video viewer right onto the twitter home page, as to reduce load times for the end user. Heck, even a pop-up system like those found on forums would be nice, but one thing’s for sure, many users would greatly benefit from not having to load another page to view a picture or video. 
5. A properly implemented search function. Although there is a search function, I find it somewhat discreet. The only link to the search function is on the bottom of profile pages, so I’m guessing that very little people know that it exists. What I’d like to see is search boxes on the right sidebar on each and every page. On profile pages, a “Search user updates” bar, on future group pages, a “Search group updates” bar, you get the idea. Then, the #tag system could actually be used for something. 

Maxime Rousseau All, Internet Internet, micro blogging, service, sms, tag, twitter

If you’ve been following hardware news for the last 2 years, you may have notice some of the many trends that are becoming the new hotness in computing. As always the race for price/performance between Intel and AMD is on, the two giants along with Nvidia are releasing new flagship products, core shrink remixes almost every month. That’s nothing new… But other hot topics are also shaking up the hardware world, namely more and more parallelized computing via multi-core CPUs, GP-GPUs and GPU acceleration, the growing presence of DDR3 in high performance systems, and something which I think is really special, the hype over smaller, cheaper, more power efficient computers, the topic of this article. 

If you look around, you’ll realized pretty much everybody is trying to get a stab at the new nettop/netdesk market: Via has been in the game for years of course, but both Intel and AMD are churning out more and more interesting products. The long awaited Atom is pretty much leading the ball, but the Athlon 2000+, and the upcoming Bobcat are also very promising alternatives. As for the actual manufacturers of such integrated systems, it’s pretty much a free for all. Asus, MSI, Acer, HTC, HP, Fujitsu, everybody wants a piece of the pie. Since I find that releasing old revamped technology is almost an insult to the thousands of engineers who have been working their asses off to make better, more powerful chips, allow me to question this new hype: tiny computing has already caught on, but is it going to live on? Will people really continue buying underpowered hardware just because it costs peanuts? 

I can see valid arguments on both side of the scale. I know that this isn’t necessarily the right place to inquire about this, but if you think about it, what applications apart from games do you use? Chances are that unless you’re doing some hardcore 3d modeling with SolidWorks or touching up video in Premiere, all you’re doing is either surfing the net, writing up documents, and chatting on AIM or MSN. Do you really need an e8600 with 8 gigs of RAM to do that? Absolutely not, and chances are that most people currently can’t afford a computer won’t be doing much more than that either, so all these mini laptops are ideal for them.

I can also see those new chips working wonders in carputers, boatputers(?), and computer aided navigation/entertainment devices of all sorts. Carputers have been around for a while, but most will agree that the Atom is pretty much the beefiest chip available in an integrated nano-ITX form factor which can fit pretty much anywhere. With the rising popularity of home servers, the integrated Atom board is also a great choice which has a smaller physical footprint, lower price and lower power consumption then the AMD Sempron powered prebuilts I keep seeing in retail stores. 

On the other hand, I’m questioning the longevity of the current low-power, low-cost chips. After all, what’s the point in giving lower income families computers if they are rendered obsolete only a few months after they are bought? The Atom 230 has been called inapt at handling day to day computing by Tom’s, which right off that bat is a pretty bad indication of how long the chip will survive, and even if it were to be good for another year or two for browsing and typing, the roadmap for the Intel Atom kind of makes me want to wait. Dual core Diamondvilles should be landing in nettops and integrated ITX devices near you any time soon, and next gen Atoms code name “Pineview” featuring a core shrink to 32nm are just around the bend, ETA 2009. Normally, I’m not the type to wait for a newer, better component because in the end you can just end up waiting eternally, but in this case, buying an Atom 230 right now is pretty much stupid.

Cost effectiveness is also an issue. Because of the almost disposable nature of the new nettops, my guess is that people will most likely want to get rid of their nettops after only a year of use (some EEE701’s are already going on sale for cheap on the net), which has me questioning if it’s better to buy a 450$ MID and having a so-so user experience for a year or shelling out under 900$ up front on a REAL laptop but enjoy it’s full sized keyboard, screen, and greater performance for 2+ years until it becomes sluggish.

Part of me just wants to empty my secondary savings account and buy a Dell Mini 9 to run OS X on, but the other part tells me that it’s a pretty stupid investment. 

So, is minimal computing really the new hotness, or is it just a passing fad? I guess that only time will tell, but in any case, I’m just happy that finally computers are becoming something that everybody can afford.

Check out the original article with replies from the Overclock.net community on my OCN blog.

Maxime Rousseau All, Hardware amd, atom, bobcat, cheap, Hardware, intel, ocn, onc, processor

Disk redundancy is something that I’ve been wanting in my home server for quite a while now, and since I now make (relatively) huge paychecks for working somewhere else than McDonalds, I thought that with the traditional HDD’s prices plumetting because of the increasinly competitive SSD segment taking over, it might be a good time to invest in some RAID1 for protection. I took a deep breath, swiped my debit care, and 200 dollars later I was the pround owner of two Seagate 7200.11 500 gig drives. 

I thought that the ICH9r southbridge on my server motherboard would make things really easy: get into the BIOS, edit a few things, press next and ok a couple of times, and maybe install a driver or two with yum to get the damned thing working. But, it turns out that RAID controllers built into motherboards kinda sucks, and that compared to a real dedicated RAID controllers, they offer no advantage versus software RAID, also called softraid by some. After some further reading, I discovered that the only real perk to chipset raid is the ability to install an OS on a RAID array without being needy of a dedicated controller. Otherwise, it’s pretty much the same thing as software RAID: since there is no dedicated RAID chip controlling where the data flows, the CPU still have to take care of all the operations concerning storage, so there are few to very little gains in terms of performance. Being to poor to afford a dedicated controller that cost in the 200$ range, I decided to go the softraid way instead. 

What I thought would be a long and tiresome process turned out to be super easy. Linux being super awesome, it has supported software RAID for quite a while, so setting up an array with a recent distro is so easy anybody with minimal skill at the command line can created their own.

First, you plug in your disks (duh), then install the actual software that creates and manages your arrays, called “mdadm”. If you’re under a Red Hat based distro or if you used yum for your software managing needs, just hit up the following at the command line.

yum install mkinitrd mdadm

After that, load a couple of modules:

modprobe raid1
modprobe linear
modprobe multipath

Note that in the previous example, I only loaded a few RAID types, because I was only planning on using RAID 1 anyways. Once that’s done, format  and partition all of your new raid disks with fdisk, as explained in this tutorial. You’ll want to use fd as partition type, which corresponds to Linux RAID autodetect. Once that is done, all you need is two commands to create your new array. 

mdadm –create /dev/md0 –level=1 –raid-disks=2 /dev/sdd /dev/sde
mdadm –examine –scan > /etc/mdadm.conf

The first command will create you array, and from then on you will have a virtual disk called md0 in your /dev/ directory, which can then be mounted wherever you wish. The second command echo’s your RAID configuration in a config file, so that your RAID array is initiated on boot. The devices list above are just the examples I used, use whatever values fit your situation. From then on, the disks will start a bit per bit sync, which according to the drive capacity and how powerful your machine is will take a while. You can monitor your array using the command “cat /proc/mdadm”. Once the sync is done, you can go ahead and do all the regular stuff you will do with a drive, like create files system on it, mount it, add it to fstab for automatic mounting, and there you go, a fully redundant array. You may have to create a filesystem on your created array at some point, which if you followed my examples and common device mounting practices will be mounted at /mnt/md0. Just google up the mkfs command, and create an ext3 filesystem on your device. 

So far, I have only run into one problem, and that was self-induced. After forgetting to save the raid array’s configuration in a file before rebooting, I attempted to recreate a similar array using /dev/sdd notation instead of /dev/sdd1. I thought I have rewritten all my drives while they were syncing, but turns out the only thing I messed up are the partition tables, and after rewriting the tables in fdisk and properly reconstructing my array, I had my data back. Big thanks to Google and this blog entry which helped me get my data back!

Now that I’m running disk redundancy, I sleep better at night thinking that my 4 year’s worth of torrenting is safe. Or not.

Maxime Rousseau All ich9, ich9r, linux, motherboard, p35, raid, seagate, Software