RAID comes in a variety of standard flavors, along with manufacturer-specific solutions (eg. X-RAID). Most of the solutions are playing very fancy tricks with your data to gracefully handle hardware (disk) failure without data loss (the exception is RAID 0). Yes, disks can fail, and when they do you typically lose all the data, so designing a system to cope with this sounds like a great idea. The marketing behind these made-for-the-home RAID devices plays to your fears of suddenly losing all our data. They tell us we need their solution to avoid a nightmare.

What they don’t tell you is their advanced RAID system adds an additional layer between you and your data. A layer which is very complex and poorly implemented. I now believe that these RAID solutions, especially the recent explosion of home NAS devices, are less reliable than the drives they replace. Over the past 3 years I’ve been using RAID (ReadyNAS), I’ve had at least five software/firmware (the RAID controller) failures, with two falsely reported drive failures. Never did any of the drives (10 total) fail. Not once. I came close to losing data multiple times, always because of poorly written software, which I really didn’t need (since the disk drive didn’t fail).

They also don’t tell you that if their software fails badly (data corrupt/missing/etc), your only choice for data recovery is sending the whole unit to them. Only them. Doing any self data repair is nearly impossible, since each system has their own proprietary solution.

Yes, everyone should plan to avoid the nightmare of a drive going up in smoke, but the answer is almost never using a RAID solution. You only need to follow one simple principle: keep at least two physical copies of all your data. That’s it. Really. Want to be super-safe, physically secure one copy (a safe/offsite/etc).

I’ve recently replaced my ReadyNAS with a set of external drives. I copy all my data to multiple physical drives. If a drive fails, I have another copy (or two!) available. Simple.

Complex systems are prone to issues – and make no mistake, World of Warcraft is an incredibly complex and ever-changing system – some downtime is to be expected

This myth is constantly thrown around as fact, even within engineering circles. Personally, I don’t expect any downtime, regardless of a system’s complexity. My digital television is fairly complex and sending hundreds of digital channels through a small wire is probably more complex, but if it stops working for even a couple minutes I’d consider it broken — and I certainly don’t buy it expecting some downtime.

What this poster (and others) probably should be saying is “we decided to design a system with some downtime to save time and money”. Continuing the myth that complex systems must be inherently flawed is destructive to the industry. People believe what they hear.

Designing a complex system to be robust is difficult, but not impossible. I just wish more people would realize this.

While the science behind what’s going on at LHC may be overwhelming for some, the quest to explore the origins of our Universe is something everyone can understand and enjoy. Now we just need to wait three months to run a 0.000000001 second experiment.

The long story:

As I was pulling into the Jamba Juice parking lot this morning, a white Dodge Neon sped diagonally across the parking lot, cut me off and parked in the space I was going to use. I pulled into the next spot instead. I guess he didn’t like that, since he had to wait for me to park before he could open his door. Guess I should have followed my usual advice and “if you see stoopid, go the other way”. While in Jamba, after I ordered, I look out the window and see he’s getting in his car, closely watching my door. I assumed he was trying to avoid hitting my car, but it turns out he was admiring his handy work.

I am often amazed at how mean people can be towards other people. I know I’ll leave this world having given more than I took, which seems to make more sense than the opposite.

To increase the open file limit for all users on pam-based Linux:

$ ulimit -n
1024
$ sudo bash
$ echo '* - nofile 65535' >> /etc/security/limits.conf
$ echo 'session required /lib/security/pam_limits.so' >> /etc/pam.d/login
# logout. login.
$ ulimit -n
65536

The LA Lakers won The NBA Championship last night, to the surprise of none. These NBA Finals were simply boring to watch. This is what the NBA has become. Having been a never-miss-a-game NBA fan throughout the 80′s and 90′s, the NBA has lost me over the past decade. It’s clear the corporate ‘make money’ attitude has overshadowed any interest they may have had in competition. Starting last year, when the Lakers surprised the league and gave The Finals to The Celtics, it was obvious the NBA wanted The Lakers to win this year.

I watched every minute of every playoff game this year. The NBA has built their game to give both sides a chance of winning. Note that this doesn’t mean the better team wins, but rather, sometimes the luckiest team, and others the chosen team. This is great for ratings (ie. money), since close games keep fans of both teams watching longer. As small as a 5% edge given to either team is usually enough to produce a victory. The Lakers get that edge just for being the most popular team. Watch every play if you don’t believe me. P.Jackson (Laker’s coach) has built a team of complaining, whinning, cry babies … and it works. They get the calls, they get to cheat the rules, they get to push the edge of the rules. The benefit is slight, but more importantly, consistent. Its enough to hand them the win.

The Lakers are a great team, and they may win without this league-given edge … but we’ll never know.

Tiffany and her new 2009 Suzuki SX4 Crossover (in Sunlight Copper Metallic)! Time for a road trip!

The Tivo Community has always been large and strong, and the Series3 users are no different. They quickly found that the Series3 units could be setup to use an external eSATA drive, extending overall capacity. This was originally intended by Tivo to be a service/developer setting, since they didn’t want to deal with the support calls, but eventually made everything plug-n-play (but still call it “unsupported”).

After some research, I found the maximum workable size for the external capacity is 1TB. Some more research pointed me towards the Fantom GreenDrive II (1TB USB2+eSATA). I ordered one from buy.com and quickly received my drive.

Installation was plug-n-play; simply power-down the Tivo, connect the drive, power-up Tivo and soon after you’re walked through the setup for the external drive. The result is 179 HD hours, a net gain of 144 HD hours!

At least now I can save some programs for more than a day and not severly affect new recordings.

Comcast setup my new digital tv/internet/phone today. I don’t really need the phone, but adding it to the package lowered my overall price. Of course, once the net access was up, the first thing I did was run DslReport’s Speed Test. Looks like both up (3Mb/s) and down (18Mb/s) speeds are at the upper-end of the range, which is a relief. There’s nothing worse than getting slow net access.

Once I get the rest of my equipment setup, I’ll connect my servers.

A loaded top-of-the-line MacPro tilts the scale at over $14,000! This is a killer system, with dual quad-core processors cranking at 2.93GHz, 32GB RAM, and 4TB of drives connected to the MacPro RAID card. But $14K?!

Apple will tell you this is not a “home computer”, but rather a Professional (ie. business) system. I suppose, if you’re looking to build a 3D rendering farm, this may be a on our shopping list.

As for me, I’ve wiped my PowerMac and getting ready to sell it (if possible?). I posted it at work, but nobody has shown any interest. I’m looking to get about $1000, which I think is a good value, especially considering the installed software. It’s actually a great system, but being PowerPC-based lowers the usefulness for most. Not sure what I’m going to do, I guess I’ll list it, with a bunch of other stuff, over the next couple weeks on craigslist. I couldn’t even think of anyone in my family that I would want to give it to.

Regardless, I think I’ll look at getting a new MacPro in about a month. No, not the $14K version.