Alde’s Stuff

If you haven’t been to one, you wouldn’t know that Mitsuwa is almost as close as someone in the United States will get to visiting Japan without actually leaving the US.

Mitsuwa is the largest Japanese supermarket in the United States (so says their website.) I’d say that’s only about half right. It’s more of a mini-mall; it has a grocery store, with produce, fresh meats, baked goods, and all the standard grocery things you’d expect in a store. Then it goes further with rows outfitted with rice cookers, beauty products, bento boxes, and the things that no good Japanese person would do without.

By half right, I’ve really only described half the store. The other half is other storefronts; a Japanese book and music store, video store, ceramic ware, and the food court. Now, an American is only used to a food court at the local mall, with the slew of fast food joints that they expect. The Mitsuwa food court is completely different. Each store is pretty unique; given that I’ve been to both the Edgewater NJ Mitsuwa and the Chicago Mitsuwa, I’ve seen that they’re different in stores.

Basically, you can get your standard Japanese quick lunch fare. One store serves ramen, another udon and soba dishes, pre-made sushi, and even Korean dishes.

This past business trip I made to Chicago, we were a scant 3 miles from Mitsuwa; we ended up eating lunch there 3 out of 4 days. The prices were adequate (the cost of 2 jumbo chicago hot dogs, a box of fries and a pop cost about the same), and the food was something you have trouble finding outside of most asian districts in big cities.

So I happened to pick up (with much restraint) 3 boxes of Pocky I’ve never seen before.

Mint Chocolate Pocky. It tastes a bit like those Andes after dinner mints, so you should definitely pick some of these up and put them in a glass so your guests can have a minty refreshment after dinner.

Also, I found Milk Chocolate Salty☆ Pocky. This was your normal pocky, with some salt embedded into the chocolate. It gave the chocolate a different taste, but shot the sodium counter through the roof.

Lastly, the oddly named Nutcream Pocky. Really, this is a creamy peanut butter pocky, instead of the normal chocolate pocky. Quite tasty.

I apologize in advance if this post is a bit ranty, but it goes to point out the fundamental flaws in the Linux desktop architecture and why Linux will never replace Windows as a true desktop OS of the masses.

The underlying problem is the graphical user interface itself, known as XOrg, X11, Xwindows, etc. Because of a drive for a “one install fits everything” model, the 25 year old software has support for hardware that hasn’t been seen for 25 years.

With requirements like that, there’s little room for monumental improvement. For example, in 2007, the world was told that Xinerama (an extension that makes multiple screens easy) was being depricated and would be replaced with RandR. However, even in the brand new X11R7.7, RandR still did not have multiple GPU support, and Xinerama is still required.

The next issue is the support for proprietary video drivers. Such support requires a multi-level approach to patching that any non-linux system administrator would easily falter and installing them. The issue is a licensing issue, not a functionality issue. It’s much easier to download and install under the covers and have an end user click a License acceptance term box than what is required to install these drivers.

I heard yesterday that Gabe Newell (of Valve/Steam/Portal/Team Fortress 2/etc) ripped Windows 8 and wants to run more games on linux.

Gabe, if you’re going to run more games on linux, you need to do something about the GUI in Linux. To do this, you’re going to have to accomplish one of 3 things:

1. Start up a 4th competing GPU company with open-sourced hardware/software that can compete with NVidia, ATI and Intel
2. Convince NVidia and ATI to truly open-source their drivers.
3. Replace XOrg/XWindows/X with something that truly can compete.

To show the difficulty here, I’m documenting my hell with trying to get a Multi-GPU Multi-Display Desktop up and running.

My desktop is a Dell Optiplex Gx 980 Tower . I have two ATI Radeon HD3450 cards installed, with 4 displays total, two per card.

For ease of use, Ubuntu 12.04LTS is the OS of choice.

First Step: Install off Ubuntu LiveDVD 12.04 downloaded from http://cdimage.ubuntu.com/releases/12.04/release/ubuntu-12.04-dvd-amd64.iso. Burned this to a DVD. Rebooted.

Choosing to Download updates while installing, and install third-party software. Then chose to install one big fat partition. Chose Timezone, English Keyboard. Created my user and computer name. Chose not to import anything from my last attempt at this; Fresh Install! Rebooting.

After Rebooting, Ubuntu/X has only put video on two of of the monitors.
I then removed network-manager, hard set my IP address the old-fashioned way, and then got to making a normal machine. Got rid up cups, which seems to be a herculean task as Ubuntu wants to install gobs of HP drivers. Then installed openssh-server so that I can connect to this machine from my laptop.

At this point in time I finally logged into the GUI front end, as all the previous work was done at the console. Except now Ubuntu wanted to send an error report back to Canonical. Seems like aptd crashed, and seems to be a known issue. Since the bug seems to be fixed in a newer version, I used updated manager to update the machine – 370 packages need updating. Because it seemed to update the kernel, I rebooted.

Task #1: Make it such that the 2 displays I have are not displaying the same thing. Using the Unity desktop, I began by looking under System Settings. There’s a nice thing there fore displays. This was simple for me to take off the screen mirroring, but it still only detected two displays.

So I decided to do what an inexperienced person might do; search thru the Ubuntu Software Center to see if there was something there for multiple GPUs. While some applications come up in an App-Store like way, the majority of packages are basically extended apt-cache information; hardly usable for a normal user.

Oh, and during this time, Compiz crashed. This was the first issue to come up due to my non-standard setup. Compiz is software that uses OpenGL to provide pretty things for GUIs (drop shadows, etc). Segfault-Crash; having to do with edge detection, so I’m assuming it was supposed to detect when the mouse pointer moved over a window to do something graphical.

Since I couldn’t find anything immediate in the settings, I ventured to click on “Additional Drivers.” To my amazement, it told me that no proprietary drivers were in use, but that I could activate the ATI/AMD FGLRX driver. So I did, with no expectations that they’d actually install. This way I was not surprised when it didn’t. “Sorry, installation of this driver failed. Please have a look at the log file for details: /var/log/jockey.log”

At this point in time, your normal desktop user bails, since /var/log/jockey.log has no meaning to them, and if they just happen to figure out how to pull up the file in a textfile viewer. So, when it spits out “DEBUG: XorgDriverHandler(%s, %s).enabled(): No X.org driver set, not checking” that barely qualifies for “Hey this didn’t work, fix it.” Of course, %s %s is a bad bad issue in any text output, seems like someone didn’t have anything to pass in a printf line, or screwed up and forgot to actually pass any variables.

Since that didn’t work, I then activated the non-post-release updates driver. This actually installed with a “You need to restart the computer to activate this driver.”

Everything Starts Going Bad…

After rebooting, X had decided to re-mirror my displays again. So, then I went back into System Settings, and un-mirrored them and hit apply. And at this point, started my long slog into the problems that is XWindows.

A display box said “The selected configuration for displays could not be applied required virtual size does not fit available size: requested=(2560, 1024), minimum=(320,200), maximum=(1280,1280)” Adding insult to injury, clicking OK gets “Failed to apply configuration: %s GDBus.Error:org.gtk.GDBus.UnmappedGError.Quark._gnome_2drr_2derror_2dquark.Code3: requested virtual size does not fit available size: requested=(2560,1024), minimum=(320,200), maximum=(1280,1280).” Note again, the %s.

So, back to Additional Drivers. Hey, the ATI/AMD FGLRX graphics driver is activated and currently in use. Maybe this time I can activate the post-release updates driver… Nope, “/sys/module/fglrx_updates/drivers does not exist, cannot rebind fglrx_updates driver”

So at this point in time, a normal user now has a broken install, and something that can’t run but a single display. Regardless of me having two GPUs, this issue affects any ATI user with dual monitors.

I then did what any normal person would do: Google!
First Link: http://askubuntu.com/questions/137251/dual-monitor-in-12-04-sort-of-works. States I need to run the Catalyst Control Center. Found it in Unity, using the search.

Started it up, Told it to Muliple-Display desktop with display(s) 2. And… Reboot. Who ever said you never had to reboot in Linux.

Back into System Settings, And ‘lo and behold, I’ve now got non-mirrored displays. Oh, and Compiz Crashed again.

So, now I’ve got two monitors, again, but not 4.

Dmesg shows that Linux at least detects my two cards:
[ 1.791232] vgaarb: bridge control possible 0000:02:00.0
[ 1.791233] vgaarb: bridge control possible 0000:01:00.0

And the fglrx driver does see both devices:
[ 13.206517] [fglrx] vendor: 1002 device: 95c5 count: 1
[ 13.206520] [fglrx] vendor: 1002 device: 95c5 count: 2

[ 13.207209] pci 0000:01:00.0: enabling device (0002 -> 0003)
[ 13.207217] pci 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[ 13.207256] [fglrx] ioport: bar 4, base 0xcc00, size: 0×100
[ 13.207263] pci 0000:02:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[ 13.207269] pci 0000:02:00.0: setting latency timer to 64

Using the Catalyst Control Center, there, next to the two display icons was a card icon. So I enabled that. This got me video on my other two monitors. System settings, Displays… Only two monitors. Back into CCC. All 4 monitors detected. Change Display properties, locations of monitors… Reboot.

I now have 4 monitors lit, but am unable to drag applications or “extend the desktop” to the other two. Can’t do it in CCC, can’t do it in Display properties. After installing and uninstalling both myunity and unsettings, I decided to go old school and see about turning Xinerama on. This was greyed out in CCC, so I figured I would hand-edit my xorg.conf file to override this.

At this point in time, I now have 4 monitors with Unity able to address all 4 of them. Easy right?

I had hooked up the scanner (along with the cassette tape player and the VCR deck) so I could digitize some of the goods / pictures / videos that my mom had collected / taken / performed in…

Things like, oh,

a Duke Ellington Program from 1966, or a video of her doing karaoke from 1991, or some of the 100s of thousands (yes, at least 6 figures of pictures) she took.

I took the time to resort some of the things she had kept for me, I found my ticket from the Star Trek 25th Anniversary Marathon… Sit Long and Propsper!

For the past couple of years, I’ve been patiently sitting on the sideline with regards to choosing a streaming media player. Yes, I know I can build my own HTPC and join the DIY crowd, however building and maintaining an HTPC is not something I want to spend my spare time doing; I’d rather pay someone to do so by buying a commercial product.

The commercial streaming market has matured and settled down in the past 2 years. Previously you had to have faith in unknown manufacturers with wonky GUIs and questionable reliability, and now you have the big internet players involved on most fronts: Google, Apple, Netflix, Western Digital, etc.

My reticence to get into the market is that my requirements haven’t changed, but have been difficulty to find in a single solution:

1. Network File Sharing: I already have a server PC with a Raid-5 array that I use for archiving the growing amount of DSLR RAW files, as well as 720p HD video taken with my Nikon D90. Whatever device must have the ability to use a network file sharing protocol that is not HTTP and does not use UPNP for file streaming.
2. Netflix: Device must be able to stream Netflix.
3. Hulu/Hulu Plus: Should be able to stream Hulu.
4. Amazon Video on Demand: My wife has gotten hooked on AVoD, and since we are now Prime members, it’s free to stream.
5. Codec Support: The device should handle AVI/MPEG/MKV/DIVX/XVID/MP3/etc. It should not be a one-trick pony that requires re-encoding on the fly.

So I set out to gather the true information to make my decision…
Read the rest of this entry »

So Spotify.. the web’s new hotness. It’s basically the old Rhapsody or old mp3.com or napster; a streaming media library but for free and with some nice social media integration.

Had a bit of an argument with a friend as to why he thought it was better… All he had was “It’s Free.” I also pointed out that Pandora was free, and limited you to an amount of hours per month…

So once I finally got my invite, I got to see what the catch was:

“The Spotify Service can be accessed (i) as an ad-supported free-to-the-user service having no monthly cap on listening hours or a cap on number of plays of a unique track during the first 6 months following creation of your Spotify account but thereafter a cap of 10 listening hours per month and a cap of 5 plays per unique track (the “Free Service”), ”

Pandora allows for 40 hours per month, however you cannot request / play specific songs. Yes, you can create an artist channel, but Pandora’s pricing paid thru to ASCAP and BMI is based on their non-specificity of playback.

Napster is $5 a month ($10 if you mobile streaming), Rhapsody is $10 a month ($15 for more than 1 mobile player). Both non-direct station-based streaming services Last.fm and Pandora both charge $3 a month, but as stated, you can’t listen to specific songs directly. If you want mobile playing, Spotify is $10 a month. So much for Free. Napster and Rhapsody have a 14 day trial; Spotify’s is basically 180 days.

So I’m less impressed by Spotify the more I find out about it. Just not “Social” enough to care what people think about my music selection.

As a reminder to myself, the proper way to clear the DNS cache on a windows machine (be it Windows 7, Vista, or XP) is to do:

ipconfig /flushdns

In some versions of Linux (Centos 5.6 confirmed), usually the name service cache daemon is installed to manage DNS cache. To clear NSCD do

/etc/init.d/nscd restart

And… If you happen to have installed bind9 as a caching name server, use:
rndc flush

Portal 2 – $30 from Amazon.

Channel 7 news Australia’s live coverage:

http://au.news.yahoo.com/video/national/watch/23188303

3 News Christchurch link:
http://www.3news.co.nz/Video/3NewsLiveStream/tabid/876/Default.aspx

I decided that the next big thing is IPv6. Good time to get ahead of the curve; in 18 months the crap is going to hit the fan; I will have highly marketable skills as I will be a precision instrument of speed and aeromatics…

The biggest problem I’ve been running into is Feature Equality, i.e. things may work for IPv4, but most programs really don’t have any or full IPv6 feature equality.

For example, I have a Juniper firewall as a home CPE device. However, running the latest JunOS on it, I cannot enter an IPv6 address on a Vlan. But, I can via the GUI. Wacky. But when I’m talking about Feature Equality, you can configure the firewall to do DHCP using IPv4, but it won’t allow you to configure DHCPv6 on the device; I guess Juniper believes SLAAC or static IP address assignment is the way of the future. And, you can’t even setup an address name on the CLI, it doesn’t know what an IPv6 address looks like:

[edit security zones security-zone trust]
admin@firewall# set address-book address foobar 2001:470:8986:1::10/128
                                                ^
invalid ip address or hostname: 2001:470:8986:1::10/128 at '2001:470:8986:1::10/128'

Oh and lord knows DHCPv6 doesn’t have problems. There’s no ipv6 routers option. None. Been talked about, but there’s no way to specifically tell an IPv6 client what its gateway is. Because we all know that it just needs to be told it’s IP subnet, it can figure out the rest… I’d hate to be on a network with Windows XP/Vista set up for Internet Access Sharing; each Windows box will announce itself as a router (i.e. Router Announcement) and you may end up with your entire LAN routing thru a tunnel on someone’s Windows desktop instead of that Juniper Firewall (Yea, I can’t set Router Announcement priority on the SRX either).

Once you get past that, and have a dual-stack or native desktop, good luck getting full featured IPv6 out of your applications.

If I want to connect to my internal CPE device to manage its configuration, I connect to https://2001:470:8986:1::1. I do not want however, want to search for information about https://2001:470:8986:1::1…

But at least you’re not the only one; Firefox doesn’t work either…



And yes, IPv6 is working:



Oh, btw: I can hit the Forward resolving host (i.e. ipv6.google.com) but can’t hit the IP address in Chrome either.

This isn’t from the point of view of a furious subscriber to a service, but that of a peer in a similar industry.

From Lord Of The Rings Online:

As the final stage of our datacenter move, all Turbine games will be offline on Tuesday February 23, from 4:00AM – 4:00PM Eastern Time (-5 GMT). Websites, including myaccount.turbine.com, forums, wikis, and social networks will be available, but players may be unable to log in or access their account information during this time. We thank you for your patience while we complete the move!

If I went to my boss and told her that “We need to do a datacenter move, which, in my consulting with every other operations team, will require a 12 hour downtime,” I think she’d laugh in my face and go talk to HR about my further employment.

I’ve worked in providing internet access, either to the business traveler, in a dial-up ISP, a web-hosting ISP, the worlds largest Tier1 network provider (at the time), and now providing internet-based services. I’m amazed when a business can provide this type of inept service to their customer base.

It just goes to show you that the real players in the internet-services space know how to build in redundancy, scale, and resilience into their product.

Redundancy: How about multiple datacenters guys? I understand the need to have centralized shards and back end database servers, but when your entire product goes offline because you’ve got a single point of failure somewhere shows that you need to push the data closer to the front-end servers.

Scale: If you’ve got at least 11 shards, that’s probably 10 too many. I understand the need to lower latency, really, I completely understand; jitter is my enemy. However, if none of these survives because your login server or front-end access servers can’t scale beyond a certain number of concurrent logged in users, you need to look at doing it differently, especially when all the graphics and all the maps and all the physics and all the etc is handled on the 11 Gig client installed on your users’ computers, essentially at the core of your shards is a long term storage database, a short term storage database, and tens of thousands of UDP updates that can be highly localized so that only the information the character would ever see will be sent their way. Maybe look at different hardware (Sun has some highly threaded servers now that can handle the amazing amount of UDP packets required should you need to handle 20,000 users with 20ms update packetization) to break out of the norm.

Resiliency: If you have single points of failure that take down your entire system, then you need to look at developing a system that allows for diminished running should that single point go down. For example, if your huge honking 32 processor Sun/Oracle database server dies, can your customers still subscribe to your service, and use your service in a normal or degraded state? Yes, somewhere something has to track that these 6 characters defeated Kranluk, but does that need to be stored centrally or can it make its way to the central DB eventually?

I interviewed for a job which had the complete change management and maintenance mode one in any operations group would dream for; at 5pm on Friday night, they shut down their service. From that time until 8am on Sunday morning, they had full reign to re-install servers, update router software, make firewall changes, etc. But it had to be up at 5pm on Sunday or millions of dollars in transactions would be lost per minute. As much as the cellphone industry has allowed other companies to provide the same level of poor service, the Financial Industry knows no such lack of service. I think Turbine should aspire more towards the loftier goal of a no outage service like those who run financial companies than be like those who manage cellular networks.