http://www.monkuniverse.com/ Discussions about computing from multiple viewpoints. en Copyright 2009 Sat, 03 Jan 2009 16:58:45 -0800 http://www.sixapart.com/movabletype/ http://blogs.law.harvard.edu/tech/rss DroboApps on the Drobo. The result: $uname -a Linux DroboShare 2.6.12.6-arm1 #1 Fri Sep 12 16:36:22 PDT 2008 armv5tejl unknown So the droboshare is actually an ARM system? Interesting. yes, you can run SSH and BASH on it, making it a good emergency system and the /bin directory is populated nicely. this also allows Jumbo Frames for Gigabit, setting the MTU at around 9000 or so. which I notice has really sped up network communications between my main box and the drobo. The developer community has not exploded like I thought it would. This may be due to the relatively low number of drobo users, but that number is getting bigger with time. I have no idea about how easy the SDK is to use, but I am tempted to take a look. Folding at home for Droboapps anyone? ]]> http://www.monkuniverse.com/2009/01/droboapps.html http://www.monkuniverse.com/2009/01/droboapps.html Linux drobo linux Sat, 03 Jan 2009 16:58:45 -0800 I have a good excuse for not updating this site in so long: I forgot how to log in. No, really. The directory structure I use for this blog is non-standard to avoid some security issues. However, I then forgot the structure. Also, I had misplaced the password for the server so I could check the paths. Whew. All solved. So anyway, posts I have planned for 2009: -Why populating 4 banks of RAM sucks. -How to figure out what happened when windows crashes. -Architecture evolution. -The joy of VAX. more later. http://www.monkuniverse.com/2009/01/so_this_is_the_new_year.html http://www.monkuniverse.com/2009/01/so_this_is_the_new_year.html Meta meta Sat, 03 Jan 2009 16:50:20 -0800 Once you get it out of the box, you'll find that your Drobo only connects to your system via USB 2.0. For exactly the same reasons that washing your car angers Aquamec the rain god, my purchase of this device is a sure sign that USB 3.0 will be out in a matter of minutes. Sure enough, I have learned via Slashdot that USB 3.0 is due out in Q2 2008. I should get a prize.]]> http://www.monkuniverse.com/2007/09/on_car_washing.html http://www.monkuniverse.com/2007/09/on_car_washing.html car washing slashdot usb Thu, 20 Sep 2007 20:23:27 -0800 SyncToy for the heavy lifting and rsync for accuracy, I began moving my data over. Thursday night, Drobo was at 0% capacity. Today, on a Saturday morning, it's at 67%. At 85% Drobo is programmed to start complaining at you, and above 90% it deliberately limits its throughput in order to nag you into upgrading its capacity. Basically, this is just a roundabout way for me to say that I bought a 4th drive from Newegg this morning. I used Junction to create a mount point (an "NTFS reparse point" in Microsoftese) from C:\Downloads to the corresponding Drobo directory. This way, I can drag files from my Desktop to a shortcut to C:\Downloads and have Explorer.exe assume that I mean "move" instead of "copy". By default, dragging a file from your C drive to an E drive would be a copy operation, which in this specific case I don't want to have happen. So I've moved most (not all) of my data to a little black box no bigger than the little black storage box where I used to keep 5.25" floppy diskettes. And already I'm nervous enough to warrant maxing out the 4 drive bays with another drive. It's been totally painless, save for a few bullsh*t SyncToy errors that rsync handled like a champ. I have some extra IDE drives laying around that I still need to migrate, and after that I expect my Drobo space consumption rate to level off logistically.]]> http://www.monkuniverse.com/2007/08/the_drobo_story_setup.html http://www.monkuniverse.com/2007/08/the_drobo_story_setup.html storage backup drobo hard drives ntfs rsync storage synctoy Sat, 11 Aug 2007 10:32:39 -0800 Drobo. A Drobo is a cool new gadget billed as a "data robot", an intelligent, scalable storage device that takes the fuss and fury out of reliable, scalable hard disk management. There are several downsides to owning a Drobo. For starters, it costs five-hundred wing-wangs and that's just for the housing device. Drives, which, in case you haven't caught on are really important when using a Drobo, cost extra. I found some nice cheap SATA drives on Newegg.com for $65 a pop. Once you get it out of the box, you'll find that your Drobo only connects to your system via USB 2.0. For exactly the same reasons that washing your car angers Aquamec the rain god, my purchase of this device is a sure sign that USB 3.0 will be out in a matter of minutes. The Drobo also only support NTFS and HFS+ partitions, so you're only going to have luck using the device on a Windows machine or a Mac. ext3 support or network connectivity is something that Data Robotics, Inc. is "working on". So with all these minuses working against it, why did I buy it? Well, let's look at the alternative. The Ultimate Perfect Forever-Free Storage Solution is going to involve buying a new system, setting up hardware or software RAID on multiple drives, and installing Linux in order to create an LVM configuration for your multiple drives, stitched together with RAID and formatted with a Big Bad filesystem like ReiserFS, JFS, or XFS that will never lock your data into some proprietary format that is going to require you to pay through the nose for the privilege of keeping your data handy. If you put your data on a system like this, it is safe for as long as all your drives work. If one drive dies, you power the machine off, replace it, and rebuild your array. This solution is scalable, open-source, and gives you unlimited options for how you shape your storage. It can be maintained forever and easily transferred to another setup five years down the road when the hardware starts to flake. This solution is also going to cost you a minimum of $500 in hardware, and you will constantly wage the risk of asking yourself "Even if I only pay $25 for a RAID controller, do I really want to entrust all of my most important data to a $25 RAID controller?" So my opinion is that you should not scrimp on your Ultimate Perfect Forever-Free Storage Solution. Trying to low-ball your hardware is only going to cause headaches at best, and data loss at worst. So even once you've taken the risk of getting some cheap gear together, you now need to assemble it. Plan on at least one evening spent building your box, putting SATA drives in, screwing them in, taking them out, turning them right-side-up, then putting them back in again. Once everything POSTs, you still have to install CentOS, fdisk the drives, set up RAID devices, set up LVM, and choose your file system. If you want to get to your data from across the network, that's going to require an NFS or Samba config, too. How much is your time worth to you? I would love, positively love, to do all of this. It's good experience and builds character. I want to be able to trust that my data is safe because I've built every last layer of protection into the system with my own two hands. But you know what? I have, at this exact moment in time, precisely 18.1 GB of free space on my C drive, and this precipitates a rather urgent need for a scalable storage solution that isn't going to cost me both $500 and an entire weekend. The Drobo, by comparison, is pretty much a "plug it in, go have a beer" kind of technology. It's pricey and it only works correctly when directly attached to a Windows system. But you know what? So are consultants.]]> http://www.monkuniverse.com/2007/08/on_storage.html http://www.monkuniverse.com/2007/08/on_storage.html storage centos drobo linux lvm sata storage usb Sat, 04 Aug 2007 14:43:22 -0800 x86 has won. Really. Yeah, your port of linux to the PS3, or the ARM, or the toaster, or your dog is neat and all, but are you really going to spend the time and effort making sure it stays maintained? This past weekend I tried getting X working on an ultrasparc 80. X.org has a wonderful bug where it can't handle how sparc's divide up PCI slots. it isn't rocket science: 0000 or 0001 as the prefix for the address. However this will be broken until (at least) X.org 7.3, which won't really be in any distros until, say, July at the earliest. If you look at the alpha releases of ubuntu, they haven't even bothered to release sparc at the same time as x86 and x64. there are a good number of sparcs out there. So I can only imagine what the story is like on rarer hardware platforms. My proposal is to consider doing what microsoft did. Microsoft NT used to be available for MIPS, PPC, and X86. It was frustrating as you had to test and maintain all of them. Finally the axe was dropped and x86 became the only survivor (yes yes, later x64 and IA64 were added, but x64 is an extension of x86 and IA64 has intel jumping up and down on its chest to keep the heart beating. Imagine all the effort and work being done on weird platforms that instead could be done on focusing on x86. You can't use the mac argument anymore as the macs are running on x86. If linux wants to be really dangerous, it needs to focus. Focus on a few key architectures. Go through freshmeat or sourceforge and kill the dying or abandoned projects. lean. mean. fearsome. consolidate your distros to 3 or 4 major ones and let the others inherit from those. http://www.monkuniverse.com/2007/06/kill_them_all.html http://www.monkuniverse.com/2007/06/kill_them_all.html Linux Wed, 27 Jun 2007 21:22:20 -0800 Today a friend of mine IM'd me asking for advice on how to bridge a MySQL database with .NET. Aside from the fact that I normally charge a significant amount of money per hour to answer questions like that, I found it interesting that he was asking me for advice on a subject I was, at that very moment, executing on behalf of a client. The client has a Microsoft SQL Server database that has some very important data, and a Linux-based MySQL server that runs a very important program. Hmmm. My first two approaches involved writing scripts to be run on the Linux server to pull information from the Windows system and push it into MySQL. Approach one was a script that would connect to the SQL Server via ODBC, pull records, and push them into MySQL. Perl, Python, Ruby, whatever. There were numerous languages that could perform this function. The key ingredient, of course, is a Linux-based ODBC driver. Sadly, ODBC drivers for Linux come in two classes: trial versions and the kind you have to pay for the privilege to use. Well, I for one refuse to call up my client and tell them "Hey, great news on the project. We have a script that can speak ODBC to the SQL Server and everything's going to be just peachy for 14 days. After that, you're going to have to buy a license for an amount of money that is undisclosed on the company's website (never a good sign). Oh, and before I forget: upgrading the OS or making a change to your hardware contractually invalidates the license and requires you to purchase another one. Oh, and the drivers won't run unless the licensing service is running, too. Toodles." Screw that. I don't mind -- too much, anyway -- paying for quality software. But to expect me to continue paying? For something I already bought? I refuse to let my vendors treat me like a criminal, or anything other than a customer who has purchased a product and now has full rights to use it in whatever manner he sees fit. I also don't buy DRM'd music. Go figure. Approach the Second was a slightly more convoluted idea: export the data from the SQL Server to a Windows share that is mountedon the Linux system via Samba. The script in use here was a simple line-by-line parser that only dealt with a comma-delimited text file and MySQL. After about three minutes of explaining approaches the First and Second to my boss, he blinked, asked me a basic question about ODBC, and then asked me if it would be possible to hook an ODBC driver directly into the export wizard in Windows. Blink. MySQL is free, and so is the ODBC driver that they provide for it. No trial versions. No "Lite Editions". MySQL offers it as a simple download on their site: you don't even need to register or create an account. I downloaded the ODBC connector, installed it, and had the DTS export configured and working on two separate systems in under 30 minutes. So, to recap: pulling from SQL Server to MySQL via ODBC? Expensive. Licenses. Scripting. Pushing from SQL Server to MySQL via ODBC? Faster. Easier. Safer. Free. To the extent that I found a better solution because I wanted to avoid the draconian licensing problems of the first, most obvious solution, I'm glad I spent an afternoon working on an alternative. I do not feel that there is anything to be gained from the idea of "lending" someone your software for a limited time and contingent on them never upgrading or changing their system in any way. http://www.monkuniverse.com/2007/05/on_licensing.html http://www.monkuniverse.com/2007/05/on_licensing.html legalese Wed, 30 May 2007 17:07:45 -0800 Get it here Later on I will show you how to turn Server 2003 into a workstation. It isn't as weird as it sounds. ]]> http://www.monkuniverse.com/2007/05/the_joy_of_server.html http://www.monkuniverse.com/2007/05/the_joy_of_server.html Windows Wed, 16 May 2007 21:24:13 -0800 http://www.monkuniverse.com/2007/05/the_tragedy_of_choices_1.html http://www.monkuniverse.com/2007/05/the_tragedy_of_choices_1.html Linux distributions iso linux Fri, 11 May 2007 21:27:19 -0800 Elatte may have been what I have been waiting for. Ubuntu with an opensolaris kernel. I had my doubts, as it was in alpha stages, but I have high hopes. The Good: Autodetects 64 or 32 bit systems. Very fast apt-get dist-upgrade works runs under vmware The Bad: limited selection of software /proc just ain't the same no lsmod/rmmod/insmod I realize the last two of the bad are pretty much no-brainers, but I figure if they made such an effort to make the system linux-like, they would make some kind of virtual redirectors or something. I was amazed at how fast the install and execution were. further experimentation is necessary. http://www.monkuniverse.com/2007/05/grande_elatte.html http://www.monkuniverse.com/2007/05/grande_elatte.html Linux Fri, 11 May 2007 18:59:44 -0800 Groklaw. What I'd like to discuss is what the hell is wrong with SCO. There is, I think, a reason why Linux is so popular. Yes, it's inspired by UNIX. Yes, it's fast and efficient and has a ton of software packages that support it. Yes, it's (usually) free. But on top of all of that, it's easy to use. Now before you spit Coke Plus all over your keyboard and monitor, you have to understand that Linux isn't meant to be easy for your mother or your brother to use. It's not meant for them to use at all. It's meant for use by developers and technically-adept users, of which there are legion. And for these mighty code and script gurus, Linux is an easy and fun place to stage all your computing operations. By contrast, SCO is still bolting bits and bobs onto an operating that was meant for use by hardcore FORTRAN geniuses. SCO's offering, in the 21st century, is essentially the same platform that your grandfather was using back in his college days. They've rebranded it, but they haven't exactly improved it. When Dennis Ritchie sits down at a Linux machine, I can only imagine that he quietly awes at how far his ideas have gone and how many bright people have solved the problems he faced back at Bell Labs. A modern Linux system has graphics, advanced browsing features, and a plethora of shells that feature, by default (gasp!) tab completion. Whenever I find myself touching a SCO server, it is with absolute disgust and disdain for how little the product has changed. Compiling anything on a SCO system is a chore because of the numerous ABIs they've introduced, changed, and revamped. Kernel compatibility technically exists, but you are hard-pressed to actually see it work. And if you try to do it better than SCO can (and doing something better than SCO can isn't even remotely hard), they'll sue your sorry ass for violating the IP that they rightfully own solely by Mandate of Checkbook. So SCO is hemorrhaging money, and they can't offer a cheap, stable, usable, or modern product. It's only a matter of time before they become so desperate to pay off the shareholders that they pass off the cursed UNIX copyright to some other company that will try to wring some value out of it. The cycle will repeat: revamp, market, slander, then sue as a last resort. Meanwhile, the rest of the world has seen the best UNIX can offer, and have already improved upon it tenfold. You can download and burn an operating system, as UNIX-y as you want it, and effortlessly do things that your grandfather only dreamed of doing on his old mainframes. Compatibility be damned: if you want to make a good product, you need to lose sight of the licensing aspect and focus on usability. If SCO made software that was usable, they might actually be able to sell it.]]> http://www.monkuniverse.com/2007/05/what_the_hell_is_wrong_with_sc.html http://www.monkuniverse.com/2007/05/what_the_hell_is_wrong_with_sc.html legalese linux sco unix Tue, 08 May 2007 15:47:02 -0800 My middle school social studies teacher frequently gave me a hard time for not living up to my potential. "Not firing on all cylinders" was how he phrased it. Today's storage buses are in the same situation. Or are they? During construction of my last system I agonized over the fact that the Western Digital Raptor drives (10K RPM, 4.6ms seek time) were only SATA-150 and not SATA-300. as a friend pointed out to me, the concern was moot. There was no way the drive could saturate the SATA-150 pipe let alone a 300 pipe. The faster seek times would more than anything else set the winner apart. Yet system builders continue to clamor for a SATA-300 Raptor. Why? Is it the planning for future expansion? I think it is more to due with the goal of eliminating as many bottlenecks as possible. I know I have felt the frustration at being hamstrung by a certain bus or interface, so it is clearly a case where emotion can trump reason. http://www.monkuniverse.com/2007/05/not_living_up_to_your_potentia.html http://www.monkuniverse.com/2007/05/not_living_up_to_your_potentia.html storage Mon, 07 May 2007 22:03:57 -0800 works. IPv4 is proven technology. IPv4 is popular. IPv6 is none of these things. Why would anyone stop using one to begin using the other?]]> http://www.monkuniverse.com/2007/05/ipv6_a_response_to_chicken_egg.html http://www.monkuniverse.com/2007/05/ipv6_a_response_to_chicken_egg.html networking ipv4 ipv6 networking sister Mon, 07 May 2007 16:16:33 -0800 Disable what you don't need (hardware) Go into your bios and turn off the things you don't need. If you are not using your parallel or serial ports, turn them off! If you aren't using your integrated modem, turn it off! However, I would suggest keeping a record of what you do turn off. That way you can have a first step in trouble shooting why a device isn't working. Remove windows components you don't need. Chances are you probably don't need MSN explorer. in windows XP open add remove programs control panel and click on add/remove windows components. Then go through and remove what you don't need. If you aren't using MSN explorer or Outlook Express, it is just taking up space and providing potentially security issues. Update your dell firmware http://support.dell.com has updates for just all about of their models. Most of the updates help performance, compatability, or reliability. Just make sure you don't lose power to the system during the update, and back up your data first. ]]> http://www.monkuniverse.com/2007/05/computer_optimization_i.html http://www.monkuniverse.com/2007/05/computer_optimization_i.html system building Sun, 06 May 2007 20:33:10 -0800 There seem to be two main camps regarding package management. Install from sources Install from binaries. This makes a big difference when one of two things: dependencies and updating. Now, I prefer the debian way in this case. apt-get install and then apt-get dist-upgrade as wanted. this works for other packages installed the same way, when it causes problems is if you compile something from source on the same system. When debian recently released 4.0, my routine updating of the system caused some problems with software that xenotrope had compiled from source. why? glibc had been updated. now with the BSD or gentoo way, this wouldn't really be a problem as the binaries would be compiled from source every time. However, this has caused me to get tired of BSD's ports system. installing the simplest package can take ages as all dependencies have to also be compiled. Now the latest version of FreeBSD claims binary installation and updating for its ports system. Arguably apt-get on debian already has this as you can specify apt-get install to install and compile source. are we seeing convergence to a standard? http://www.monkuniverse.com/2007/05/package_management_vs_optimiza.html http://www.monkuniverse.com/2007/05/package_management_vs_optimiza.html Linux Sun, 06 May 2007 20:18:52 -0800