Théoden's Coding Tips

I was initially excited at the chance to install and configure a TalkSwitch 240vs PBX phone system. With twelve TalkSwitch 850i cordless phones spread across four base stations, I figured I would have my work empire bathed in telephony goodness. And the price was very low for an IP phone system… too low, as it turns out.

FLAW I: the TalkSwitch Management Software runs on Windows only. This isn’t too big of a deal since I can use VirtualBox on Linux or Parallels on Mac, but the PC-only requirement is still a hassle. I had originally hoped that configuration could be accomplished via a web interface, but apparently not.

FLAW II: conference calling is only 3-way. You can’t have 4+ 850i’s on the same call, nor can you have more than two 850i’s talking to the same outside line. I just assumed that, like a traditional phone system, any number of phones could pick up on an outside line simultaneously. To TalkSwitch’s credit, the system isn’t designed that way because you don’t want people to be able to eavesdrop on conversations. That’s fine, except we need a way to have a “conference” that includes more than three people. You could get around that with a speakerphone, but…

FLAW III: the speakerphone is choppy and cuts in and out due to an overly sensitive mic. This was revealed to be a known flaw when I talked to tech support on the phone. This renders the speakerphone useless, and also takes away a means of working around the 3-way conference limitation (i.e., by having people sit around a speakerphone).

FLAW IV: tech support is painfully slow. I was having trouble updating the firmware on the 850i’s so I filled out a support ticket. After five days I got this response:

This issue would be best dealt with over the phone. Please contact us at 866-393-9960 option 3.

Of course option 3 wasn’t the correct option, and sometimes the tech support line sends you to voicemail instead of queuing you up. Support is not 15-30 minutes away – you’ll wait hours to get any help.

FLAW V: this one is personal. An online vendor who shall rename nameless (because they’ve been decent to me) sold me these phones. They will not take the TalkSwitch system back because I’m past the 30-day return policy limit. But the vendor did contact TalkSwitch and got their VP of sales to call me. I assumed that the call would result in TalkSwitch placating me in some way, even if just some free headsets or something. Instead, the guy said basically nothing and simply confirmed that, yes, conferencing is limited to 3-way, and yes, the speakerphones suck. He was being so useless that at one point I just blurted out, “So… why are you calling me? Are calling to console me?” What a terrible waste of time and money this has been. I thought I’d done my homework on this system before I bought it, but obviously I didn’t dig deep enough. Hopefully this post will help any IT brethren who are considering TalkSwitch; if you’re thinking about buying TalkSwitch, think again.

I recently upgraded my 3ware 9550SX-8LP RAID5 array from 250 GB drives to 1 TB drives. After the lengthy RAID initialization process, I tried to create a partition using fdisk. Unfortunately, fdisk seemed to only create a partition size of about 400 GB instead of the 6+ TB that I expected. For reference:

• fdisk version: v2.12q
• OS: SuSE Linux 9.3 (x86-64)
• kernel: 2.6.11.4-21.14-smp

After searching around for awhile, I came across Linux Creating a Partition Size Larger than 2TB by Vivek Gite. The article recommended using parted; my system had GNU Parted 1.6.21 and it worked just fine. Do remember that you need to mklabel gpt. And now df -h lists the partition at a hearty “6.4 T”.

For some reason I have the urge to do a little jig and sing “Mo’ money, mo’ money, mo’!”

Following a power outage that outlasted our UPS’s, I noticed that two out of the three modules (MIN-6250P) in one of our server’s EMACS R3U-6460P power supply were dead. And the fan on the one module that still had an active LED was not spinning. Switching out the fans from the dead modules, I found that all the fans (SUNON KDE1204PKVX) were not functional. The working module was painfully hot when I removed it from the machine – no doubt due to the lack of heat dissipation.

So in terms of the modules dying, the following progression makes sense to me: 1) fan dies, 2) module overheats, 3) some critical component melts/fries.

However, I have a hard time believing that the power outage somehow fried the fans. More likely is that the fans were already dead or dying, and perhaps the power outage and loss of A/C were the straw that broke the camel’s back?

I am going to buy a new fan for the working module, then see if it will spin up. If it works, then that would imply that the fans are just kind of shoddy – even though they are around 5 years old. If the new fan does not spin up, then I would have to assume that somehow the power outage wreaked havoc with the modules.

UPDATE (2011-02-28): finally got around to installing a new fan in the working module; it began whirring away with no problems. So with a little splicing I’ve managed to rescue this $150 power module.

The 2000s have been the first decade where I’ve actually had some serious purchasing power. While I rarely splurge, I have managed to accumulate some excellent technology over the past ten years. Here are my favorites from each year:

2009… Synology DS209j NAS (two 640 GB drives in RAID1)

Like the DreamHost of file servers, this network attached storage (NAS) device provides an intuitive, clean interface to most of the things you’d want to do with a Linux file server. No need to install software, tweak config files, or make sure that services start on boot – the Synology pretty much Just Works. I bought the DS209j and (separately) two 640 GB drives for a reasonable total of $360. I have a similar model that I use on the job and absolutely love it for that purpose.

2008… Apple MacBook Pro (15″ screen, 2.4 GHz Core 2 Duo, 2 GB RAM)

I have never been happier with any computer. Of course I’ve never spent more on any computer… this one cost me a grand total of $2,150. I later upgraded to 4 GB of RAM for a trivial sum. In terms of my computing history, I ran Windows all the way through my undergrad, then switched to Linux for the next seven or so years. Now that I’m using OS X, I have all the polish of an Apple product, with the stability and hackability of a UNIX system.

2007… Panasonic TH-50PX77U (50″ plasma, 720p)

I researched this purchase exhaustively in the AVS Forums and wound up with a good deal on an awesome television – $1,900 including shipping. Of course, these days I could get a 1080p set for less, but let’s not think about that. I expect this plasma to act as my primary television for years to come.

2006… Nintendo DS Lite

As you can see from Part I and Part II of my Nintendo DS Roundup series, I have logged many hours on this slick handheld. I love the dual screens, even for games that have no touchscreen functionality at all. The battery life is awesome, the suspend/hibernate mode works flawlessly every time I close and re-open the clamshell, and the two speakers do actually provide a decent audio experience. It’s difficult to imagine a better portable gaming system than this.

2005… PlayStation 2 (slimline)

Yes, I was certainly a latecomer to the PS2. The positive of this was that I could sift through the huge library of existing games to pull out the gems:

• Final Fantasy X
• ICO
• Shadow of the Colossus
• Resident Evil 4
• Katamari Damacy
• Final Fantasy XII

I didn’t pour my soul into this system like I did with the Gamecube, but the PS2 provided a reliable, quality console experience while also serving as a DVD player for my second television.

2004… Game Boy Advance SP

This little clamshell was a fitting ancestor to the DS. A few of my favorite games for the system:

• Mario Golf
• Advance Wars
• The Lord of the Rings: The Third Age

2003… TBD

2002… Nintendo Gamecube

I doubt that I will ever get more mileage out of a gaming system than I did with the Gamecube. Looking back at my reviews, I see a slew of great games that held me captive:

• Super Smash Bros Melee
• Eternal Darkness
• Metroid Prime
• Super Mario Sunshine
• Gladius
• Resident Evil
• Skies of Arcadia Legends
• The Legend of Zelda: the Wind Waker
• XIII
• Beyond Good & Evil
• Prince of Persia: The Sands of Time
• Paper Mario: The Thousand-Year Door

Hell, I even wrote a research paper on the Gamecube’s CPU for a computer architecture class in grad school. The wireless wavebird controllers were a brilliant innovation – especially for a cat owner. The Gamecube was there for me at a point in my life when I needed a lot of distance from certain personal matters; this is technology that has special meaning.

2001… Toyota Camry

Although I bought the car used in 2002, the Camry itself is a 2001 model. With only 24,000 miles on the odometer, I felt that the total cost of $15,200 was a great deal; my interest rate was pretty good, but let’s not think about the extra I paid to the bank. Anyway, this car is never going to get me any chicks, but it does work flawlessly, doesn’t leave me stranded, and has been very reasonable in terms of maintenance costs. I’m not a car guy, but when a machine with that many moving parts functions so reliably, you know you’re dealing with quality engineering.

2000… TBD

You might have noticed that there are no cell phones on here; having owned a Palm Pilot, Treo 650, and Centro, I’ve felt underwhelmed in that department. However, I am hoping that the Droid will find a home on my list for the next decade…

After a user complained that a 28-page PowerPoint document was taking an hour to print on our networked Lexmark C522, I narrowed the problem down to OS X 10.4 (in my case, 10.4.11).

Solution: use this PPD. In other words:

1. Save the PPD to your desktop
2. Go into System Preferences and remove the printer
3. Add the printer back, specifying the PPD on your desktop as the driver instead of whatever OS X wants to use

My understanding is that the PPD I’ve linked to is supplied by Lexmark – perhaps they simply haven’t updated the 10.4.x driver on their website.

I have two 500GB IDE drives that I will use to create a software RAID1 (mirrored) array. How should I connect them to maximize performance?

I’ve used software RAID on my personal file servers for years and ask myself that question every time I make changes. Google has never given me a satisfactory answer, probably because there are so many variables involved that the answer is too system-dependent. Even defining “performance” itself can be a little tricky.

This time I am going to do the work and figure out the best configuration for myself. To start with, here are my system vitals:

• AMD Athlon XP 1700+, 256MB RAM
• IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
• Mass storage controller: Promise Technology, Inc. PDC20268 (Ultra100 TX2) (rev 02)
• openSUSE 11.1, kernel-pae-2.6.27.19-3.2.1
• RAID1 file system: reiserfs-3.6.19-116.62
• Two WD5000AAKB HDs (EIDE, 500 GB, 100 MB/s, 16 MB Cache, 7200 RPM)

So here are all the places I could potentially connect the two drives:

I used bonnie++ -x 8 -u root and averaged the results to measure performance. I only examined configurations that were interesting to me, excluding “Pri M” because that is the location of the root drive. Here is what I found:

¹ Block Sequential Output (put_block)
² Block Sequential Input (get_block)
³ Random Seeks (seeks)

The most obvious thing that I glean from this is that writing to the software RAID becomes faster when 1) the drives are on separate channels, and 2) the drives are on the IDE bus instead of the PCI bus. Similarly, random seeks per second improve significantly when the drives are on separate channels. This all makes sense to me – hardware configurations that increase the opportunity for parallel operations are more efficient.

However, there’s one thing I don’t really understand: if the Ultra100 TX2 is capable of the same read performance as the main IDE channels, why is it not capable of the same write performance?

In the end, I think I’ll go with the {Sec M, Sec S} configuration. Here’s why:

• Although {Sec M, Pri S} offers the best performance, I don’t want to clog up both IDE channels with RAID traffic.
• Despite this obsessive performance analysis, the server is question is basically just a personal jukebox – I really don’t need screaming speeds.

There. I feel better now.