Dataman in LasVegas

SQL Server Management Studio Setting Top X

2012-02-03T09:56:00.000-08:00

Something that's bugged me about SQL 2008 (R2) is the limits expressed when you try to browse or edit all for a given table. Under Sql Server Mangement Studio (SSMS) you get this menu by right clicking on a table:

I know the argument, why should you return all the rows if you're just interested in the top 200 or so? Well, we have 350 employees in one table. If you accept the default Edit Top 200, then you are going to miss some. Typically I like to open the full table, find what I need to change, and change it. Top 200 just doesn't cut it.

I dug and found the settings for this, Tools | Options | Sql Server Object Explorer | Commands. Simply setting both of these values to 0 (zero) gives you back browsing and editing all rows in SSMS:

Anyhow, one of the purposes of this blog is offloaded memory storage. Easily searchable in case I forget. And I know I'll be looking for this one again.

ASP.NET Menu broken in IE8 Standards Mode

2012-01-09T14:19:00.000-08:00

Our enterprise updated security settings to force IE8 into Standards Based mode using Group Policy. This immediately caused several of our ASP.NET apps which depend on dynamic menus to fail. Rather than rendering a submenu, a blank white box is displayed:

I tried a hotfix, KB981201 (which superceeded KB962351), but no joy. Patch did not appear to work against Dot Net 2.0 SP2, though the package install did report success.

Scouting around, I did find a great blog article on the issue:
Kudos and thank you Mark Hildreth.

To boil it down for my tiny brain to understand:
(Red text is what I added to existing code.)

<head runat="server">
<title>eTime</title>
<style type="text/css">
.style1 { background-color: #FF0000; }
.adjustedZIndex { z-index:1; }
</style>
…
<asp:Menu …
<DynamicMenuStyle CssClass="adjustedZIndex" BackColor="#FFFBD6"/>
… /asp:menu>

I have a ticket open with Microsoft to see if they have a working patch for Dot Net 2.0 SP2. But this work around does appear to be valid, especially if you used master pages.

Beaglebone Angstrom Ports

2011-12-31T14:30:00.000-08:00

Using beaglebone and trying to test access methods. I've had to rediscover the information a few times, so putting it here for future reference.

Function	Protocol	Port	Sample use
Shell Access	SSH	22	Use putty
Slide Show Readme	HTTP	80	http://192,168.0.115
Gate1 Shell Access	HTTPS	443	https://192,168.0.115
Cloud9 Access	HTTP	3000	http://192,168.0.115:3000

Beaglebone tech document

2011-12-30T11:22:00.000-08:00

I was hunting around the Internet for more technical information on the beaglebone. I found that DigiKey partnered with TI on this device, and there was a good references on the DigiKey product page:

http://beagleboard.org/static/beaglebone/a3/Docs/Hardware/BONE_SRM.pdf

That's pretty much what i've been looking for. Pinouts, schematics and such. The document also states that it's conveniently included in the default distribution. When you connect to beaglebone via USB, the sd card is initially mounted as storage device. The tech docs are stored in the /documents/hardware folder.

Upgrading BeagleBone to Angstrom December 2012 Release

2011-12-30T01:06:00.000-08:00

I've been working on upgrading my Beagbleboard. When I checked today, I noticed that the principal distribution for Angstrom (http://www.angstrom-distribution.org/demo/beaglebone/) has a new version:

Angstrom-Cloud9-IDE-eglibc-ipk-v2011.12-core-beaglebone-2011.12.26.img.gz

To upgrade to this version, I did the following from Ubuntu 11.10:

1) Inserted Micro SD into SD holder, and inserted into computer.

2) Used Ubuntu Disk Utility to unmount and delete all volumes on card.
3) Executed: sudo sh makecard.txt /dev/sdb

4) Executed: zcat Angstrom-Cloud9-IDE-eglibc-ipk-v2011.12-core-beaglebone-2011.12.26.img.gz > /dev/sdb

5) Ejected card from Ubuntu.

6) Loaded micro SD into beaglebone, and plugged into the ethernet and power.

7) Logged onto DLINK router at 192.168.0.1

8) Observed Status | Log

Time	Message
Fri Dec 30 00:55:06 2011	DHCP: Server sending ACK to 192.168.0.115. (Lease time = 604800)
Fri Dec 30 00:55:06 2011	DHCP: Server receive REQUEST from d4:94:a1:39:59:ac.
Fri Dec 30 00:55:06 2011	DHCP: Server sending OFFER of 192.168.0.115.
Fri Dec 30 00:55:06 2011	DHCP: Server receive DISCOVER from d4:94:a1:39:59:ac.

9) http://192.168.0.115 reveals Beaglebone startup page.

Fixing up orphaned users in Sql Server

2011-12-08T08:41:00.000-08:00

As we all know, SQL Server tracks Sql Server Logins in the master table. This can sometime cause a disconnect when restoring a database from a backup taken on another server.

The symptom is that the user is unable to connect to the database properly. Looking at the database, all the proper rights are assigned, but the user still can't operate in the database.

The solution is:
use [database]
exec sp_change_users_login 'Auto_fix', '[user]', null,'[password]'

Where [database] is the database affected.
[user] is the user to fix.
[password] is the password to use for the login if the user id does not exist as a Sql Server login.

You should see a status message similar to the following:

The row for user '[user]' will be fixed by updating its login link to a login already in existence.
The number of orphaned users fixed by updating users was 1.
The number of orphaned users fixed by adding new logins and then updating users was 0

Getting BeagleBone to work with Ubuntu 11.10

2011-12-06T16:45:00.000-08:00

My BeagleBone finally arrived, and I've been having a devil of a time getting it to work due to simple misread of the instructions.

I started with a fresh copy of Windows 7, and used Wubi to install a fresh copy of Ubuntu 11.10. More info on that at Ubuntu.Com. But simply put, the installer builds a copy of the Ubuntu file system in windows, and simply adds a command in the boot manager to point to it. Best of both worlds, a version of ubuntu that you can boot into without permanently damaging your system or partitions. Once up and running, I was safe to proceed.

First and foremost, the help documents on the supplied microsd card are somewhat misleading and just plain wrong. The diagrams provided are not for the beaglebone but for the beagleboard.

A better resource is here:
http://beagleboard.org/static/beaglebone/a3/README.htm

With this, I learned that the image may have been out of date. I downloaded the current distribution (from the instructions above), and loaded it to the card as specified.

$ sudo -s
   (type in your password)
  # zcat Angstrom-Cloud9-IDE-eglibc-ipk-v2011.11-core-
beaglebone-2011.11.16.img.gz > /dev/sdb
  # exit

Next, I started the USB to Serial Driver:

sudo modprobe ftdi_sio vendor=0x0403 product=0xa6d0

Next, I was able to view bootup with:

# sudo apt-get install screen

# screen /dev/`dmesg  | grep FTDI | grep "now attached to" | 
tail -n 1 | awk '{ print $NF }'` 115200

There's a misprint in the original instructions, reads sreen rather than screen. I put the modprobe and screen commands into a script I call BeagleTerm.sh. I then marked that as executable.

Next, and here's where I got really stupid. The instructions say to eject the microsd card. So, duh, I pushed on and ejected the microsd against my better judgment. That's the duh moment. They don't want you to eject the microsd from the beaglebone, they want you to eject it from your computer. So, using folders, click eject.

The beaglebone will automatically configure to 192.168.7.2

http://192.168.7.2 is a presentation
http://192.168.7.2:3000 is the onboard webrowser based IDE.

Still no SG1 on Apple Tv

2011-12-05T01:32:00.001-08:00

Stargate SG1 has been broken on Apple TV for more than a week now. According to tech support at Netflix, they have been unable to track down the problem. To me, it's looking more and more like an Apple TV problem. SG1 streams properly from Xbox, pc, and even iPad. I've even been able to AirPlay from iPod to Apple TV. So why not directly from Aplle TV. We still don't know why.

Sql Server Methods of Querying Remote Data

2011-08-29T14:52:00.000-07:00

There are four major methods for querying remote data from TSQL in SQL Server 2008 R2.

This article looks quickly at all 4 methods:

From a 4 Part Naming Convention.

From OpenQuery

From OpenRowSet

From OpenDataSource

Yes, Yes, Integation Services has all sorts of cool remote data features. In this article we are primarily concerned about what we can do directly in TSQL.

4 Part Naming Convention:
The 4 part naming convention allows use of remote tables as if they were local tables. The caveat is that the linked data source must have bee previously configured. (See http://datamanlv.blogspot.com/2011/04/sql-2008-r2-64-talking-to-oracle-11g-64.html) Even if configured, not all vendors support the 4 part naming convention. It seems to work best wth SQL Server based linked servers.

The basics of the convention are [Server].[Database].[Schema].[Table]. Quotes are not used.

Server is one of your previously configured linked servers, without quotes. Example SQL02.
Database is the database name on the remote server. Example eTime.
Schema is typically dbo, but may be skipped entirely. Example dbo.
Table is the table being accessed from withing the schema. Example Employees.

A few examples of using the 4 part naming convention:
Select * from SQL02.eTime.dbo.Employees
Select * from SQL02.eTime..Employees
Both syntactically identical.

In the case that the desired table is in another database on the same instance that the query is being run from, the database term may be dropped:
Use eLearning
Select * from eTime.dbo.Employees

Nice part about this is you can use the table as if it were local, join to it, select, whatever you need.
Select * from Requests R inner join eTime.dbo.Employee E on R.EmployeeID = E.EmployeeID

The language is quite natural but does have a few pitfalls. The most notable is with odd server names or instances. In this case, it's best to surround them in brackets [ ].
Select * from [WebPublic\SqlExpress].eTime.dbo.Employees

Unfortunately, not all servers support the 4 part naming convention. Most notable Oracle 10g doesn't seem to like it. Enter the next 2 options:

OpenQuery:

The OpenQuery command lets you use pass through queries inline as if they were tables. Easiest way to see this is with an example:

Select * from OpenQuery(FINP,'SELECT * FROM APPS.EMPLOYEES')

This command will simply pass the text query through the server. You can join local tables to the returned data, or whatever you like.

Select * from Requests R inner join OpenQuery(FINP,'SELECT * FROM APPS.EMPLOYEES') E
on R.EmployeeID = E.EmployeeID

It's not terribly efficient, and will probably fetch the entire Oracle Query locally before doing the join.

Note that the string is non negotiably static. You can't have fields or any kind of parameters in the string. Yes, that does make it a lot less useful. But there is a workaround....
DECLARE @SQL VarChar(8000)
SET @SQL = 'SELECT * FROM OpenQuery(FINP,''SELECT * FROM APPS.REQUESTS WHERE EMPNO = ' + LTRIM(STR(EMPNO)) + ''''
EXEC (@SQL)

You could even insert the return into a temp table...
DECLARE #Table Table (EmployeeNumber INT)

INSERT INTO #Table
EXEC (@SQL)

Note that Adhoc Queries must be enabled on server, and you must use the ( )'s with EXEC.

OpenRowSet:

Is a lot like OpenQuery, except you do not need to the previously define the connection. You can basically define the linked server on the fly. While this exposes passwords in your SQL Script, it does protect you from creating a linked server that someone counld possibly use.

SELECT * FROM OpenRowSet('MSDAORA','SERVER.WORLD';'USER';'PASS','SELECT * FROM TABLE')

Note the odd syntax for the second parameter. That's 'SERVER';'USERID';'PASSWORD' with semicolons between the indivicual strings.

This command is subject to the same no funny business with variables as OpenQuery. I've used the same bypass successfully.

OpenDataSource:

Lastly, OpenDataSource works on the fly much like OpenRowSet. You don't need to predefine, but in the case of this command, you accept the full contents of the indicated table or view.

SELECT * FROM OPENDATASOURCE('SQLNCLI','Data Source=London\Payroll;Integrated Security=SSPI').AdventureWorks.HumanResources.Employee

Basically, it's the linked server, without having to first create the server. SQLNCLI is also a nifty trick for using the current Dot Net Connector, whatever it may be really called.

I've seen this more typically used with Excel files...

SELECT * FROM OPENDATASOURCE('Microsoft.Jet.OLEDB.4.0',
'Data Source=C:\DataFolder\Documents\TestExcel.xls;Extended Properties=EXCEL 5.0')...[Sheet1$] ;

Summary:

Each of these methods have their uses and places.

4 Part Naming Convention: The easiest and simplest way to link to the other allows instances.
OpenQuery: Best if your datasource is a predefined linked server that does not allow 4 part.
OpenRowSet / OpenDataSource: Best if you need to define your connection on the fly.

From my brain, experience, and books online.

Converting a Dynamic Disk back to Basic Windows Server 2003

2011-08-24T13:09:00.000-07:00

We had a SQL Server that had gone space critical on C: drive due to poor initial sizing. Poor thing was working with less than 800mb of disk space afer all the space savings tricks I could muster. The only choice was to resize or rebuild the server. Adding to the complication, it's a boot from SAN server. And at some point in the past, the initially basic boot drive was converted to dynamic.

Gparted under Ubunutu 11.04 was unable to resize the volume as it does not work with dynamic volumes. After some research, I was able to develop a process to convert the disk back to basic with data in place.
dskProbe can be downloaded at:
http://www.microsoft.com/download/en/details.aspx?displaylang=en&id=18546
Would not install directly on Server 2003.  Must be installed on XP, but can easily be copied to Server 2003.

This process only works if you have not used ANY of the features of dyanmic disks. Which isn't difficult as boot disks which are dynamic but formerly basic, can't be spanned, at least under 2003.

1) Clone C: Partition

2) dskProbe (A Microsoft XP SP2 utility)
   a. Drives |   Pyshical Drives… | Double click Physical Drive 0 | Click Handle 0 Set Active
     OK
    b. Sectors | Read | Read
    c. At address 1C2, change 42 to 07.
    d. At address 1D2, change 42 to 07.
        (Possibly higher address/es as well)
    e. Sectors | Write | Yes | Write It | Yes
    f. File | Exit
3) Reboot
4) Start | RUN | CMD | chkdsk
     If error: chkdsk / f
5) Test
6) Delete Clone

As part of testing this process,
I neeed to convert the boot volume to dynamic.
This is the procedure I used:

1) Start | RUN | CMD
2) Diskpart
3) List Disk
4) Select Disk 0
5) Convert Dynamic
6) Exit
7) Exit
8) Reboot (Required 2 reboots)

It's official, I'm a geek...

2011-08-24T12:55:00.000-07:00

My car was recently in a geeky license plate slide show on Network World. I guess I'm officially a geek. More plates here: http://www.networkworld.com/slideshows/2011/081711-geeky-license-plates.html?source=NWWNLE_nlt_daily_pm_2011-08-23#slide48

Resizing a VMWare Drive for Server 2003

2011-08-11T09:32:00.000-07:00

We have a number of poorly sized C drives in our Virtual IT Center. Admins have been reluctant to resize the partitions are there is no native support in Server 2003 for resizing the C drive. I have been able to successfully resize C drives on Server 2003 (both physical and virtual) using the following method:

1) Shut down the VM
2) Make a backup of the VM and partions.
3) Add storage to the existing partion using VMWare.
4) Mount an Ubutnu 11.04 ISO and boot to it.
5) Run GPARTED resizing the partion to consume the new space.
6) Shut down and eject the ISO.
7) Restart Windows
8) Don't be alarmed!
    An automatic chkdsk will be executed,
    The drive fixed up,
    And windows will then boot normally.

MS SQL connection to Oracle on the fly

2011-08-10T14:32:00.000-07:00

For security reasons, I'd like to connect to Oracle on the fly from Sql Server. That's connecting to Oracle without first creating a linked server.

SELECT * FROM

OpenRowSet('MSDAORA','SERVER.WORLD';'USER';'PASS','SELECT * FROM TABLE')

Note the odd syntax for the second parameter:
'SERVER.WORLD';'USER';'PASS'
Where Server.World is the Instance to connect to.
User is the username,
Pass is the password.

Connecting to MYSQL from ASP.NET

2011-08-08T19:37:00.000-07:00

Ok, I'm a little behind the curve on this one,
But I wanted to connect to MySQL from ASP.NET 2.0
That's the best I can do with my cheap hosting account on Brinkster.Com

I was looking for a concise example,
And ran into a couple of a problems.
Turns out the current mysql dot net connector is buggy and doesn't work.
Had to end up using a prior version, 6.37
http://dev.mysql.com/downloads/connector/net/6.3.html

For the rest of the setup,
I used Microsoft Visual Studio Express 2010
Which you can automatically install using Microsoft Web Platform Installer 3.0
http://www.microsoft.com/web/downloads/platform.aspx
Just select Visual Web Developer Express 2010
The installer will automatically install all other required (and not required) components.

Start with:
New WebSite
Visual Basic | Web | ASP.Net Empty Web Application

Back to Windows,
Navigate to: C:\Program Files (x86)\MySQL\MySQL Connector Net 6.3.7\Assemblies\v2.0
Copy: MySql.Data.Dll
Paste to: Project BIN folder

Back to project,
Right click project,
Add Reference | BIN Tab | Navigate to Bin | Select MySql.Data.DLL

Right click project again,

Add | New Item | Visual Basic | Web | Web Form |  Name as Default.ASPX
Save the project
Open Default.ASPX
Right Click, View Code
Replace all vb code with the following:
Update the x's in connection string.
And run!

Imports System.Data

Imports MySql.Data.MySqlClient
Partial Class _Default
Inherits System.Web.UI.Page
Protected Sub Page_Load(sender As Object, e As System.EventArgs) Handles Me.Load

Try
   Dim da As New MySqlDataAdapter("Select * from Table", _
   "Server=x.com;Database=x;user id=x;password=x;pooling=false")
   Dim ds As New DataSet
   da.Fill(ds)
   Dim dg As New DataGrid()
   dg.DataSource = ds
   Page.Controls.Add(dg)
   Page.DataBind()
Catch ex As Exception
   Response.Write("Error: " & ex.ToString)
End Try
End Sub
End Class

Cleanly Killing SQL Processes

2011-07-18T11:21:00.000-07:00

And speaking of executions!
At work we have Industrial SQL, INSQL, which is an archiving database engine that's attached to SQL Server. When we run bad queries in SQL, sometimes the INSQL side hangs and we need to kill the jobs.

So what's really happening here is that SQL executes the command, and waits for a separate Windows process (INSQL) to finish. Killing the SQL command does not kill the Windows task, which must be killed manually.

To that end, you can review currently active tasks with:
SELECT * FROM MASTER..SysProcesses
WHERE LOGINNAME='X'

That will show you all the tasks associate with a login,
Find the one with the high wait time,
As such:

spid | kpid | blocked | waittype | waittime
108 | 3312 | 0 | 0x006D | 5151406

So, execute KILL SPID to kill the SQL Task.
KILL 108

Use Windows Task Manager to kill task KPID.
TASKKILL /PID 3312

Be careful when naming reports!

2011-07-06T11:33:00.000-07:00

We had an interesting situation today...

One of the reports on our Sql Reporting Services Server is named, innocently enough, Maximo Supervisor. It's a report that shows the current manager assigned to each department in a popular piece of software. The problem was, one of the user subscribed to the report, and this is what she recieved this morning:

Maximo Supervisor was executed at 7/4/2011 2:00:06 AM

Well, you know how sensative those users can be about these things.... ;-)

While it used to be company policy to execute the occasional Maximo Supervisor to boost moral, apparently the practice is now frowned upon. From now on we'll only be executing the Maximo Supervisor Report.

SQL 2008 R2 Reporting Services SSRS fails to follow mirrored database

2011-06-28T13:48:00.000-07:00

I'm working with a dbMinion, Ed Callanta, who's been trying to track down a reporting services failover problem for me. I have three servers set for mirroring, A, B, and W. A is the Primary, B is the Mirror and W is the Witness. I have both a report and dot net 4 application that reference the eLog database on A. Connection strings for both report and dot net app were set as follows:
Reporting Services:
Data Source=A;Failover Partner=B;Initial Catalog=eLog;Pooling=False

Dot Net App:
Data Source=A;Failover Partner=B;Initial Catalog=eLog;User=X;Password=X;Pooling=False

Ok, so start the Dot Net app, enter some data.
Observe, data recorded to db.
Fail eLog over to B, enter some data.
Observe, data recorded to db.
Fail back to A, enter some data.
Observe, data recorded to db.

Now, lets try reporting services.
With eLog on A, run the report.
Observe, report runs correctly.
Fail eLog over to B, run the report.
Fails!

An error has occured during report processing. (rsProcessAborted)
Query execution failed for database 'dsDefaultLog'. (rsErrorExecutingCommand)
For more information about this error navigate to the report server on the local server machine, or enable remote errors.

Ok, I'm game, what's wrong with Reporting Services.

DUH, Ed found it, nothing wrong with Reporting Services.
Read the error message, the problem is with the dsDefaultLog recordset.
Open the report and check the query for dsDefaultLog:

Select GroupID, Supervisorlevel
from etime.dbo.Supervisors ...

Well, if only eLog failed to B,
And eTime is still on A,
How is eLog supposed to be able to find eTime?

This seems like a gap in the failover logic.
For Sql Server assumes that eTime is on the same server that we are currently executing from.
In the case of failing one db only, eLog, eTime did not move as well.
Can we tell SQL server where to find eTime as well, in a fault tollerant method?

Configure dbMail hangs in SQL Server 2008

2011-06-27T14:20:00.000-07:00

I've had a 2008 Server on my network that I've just not been able to complete dbMail configuration on. Every time I started, the process would hang trying to enable Brokerage Service. So, this has always been a non-emailing stepchild on the network. Turns out there's a known resource contention lock between dbMail configuration and Sql Server Agent. Stopping agent seems to do the trick. dbMail configuration gets past the block almost immediately. You can restart the agent at that point.

What a difference a little debugging makes...

2011-06-12T00:52:00.000-07:00

I've been working with Joey of LookWhatJoeysMaking.Com. The kid certainly has potential. Tonight he was using old camera guts to make high intensity sparks. He's scheming on building a Jacob's Ladder no doubt. Asking all the right questions, and I've been advising him on safety. The spark is certainly alive in this one.

We've been working on a tutorial for the 3 Asix Accelerometer from Modern Devices. It's been a while since I made a scrolling image in VB, so nows a good time.

This was my first shot at a plotter for a X Red, Y Green, Z Blue..

And after a bit of debugging,
I was able to clean up the image...

Met this awesome kid...

2011-06-10T22:11:00.000-07:00

Ran into this awesome kid at Makerfair. Joey of LookWhatJoeysMaking.Com is a 14 year old that's already designed and built a marshmellow canon (now why didn't I think of that?) And is just starting to tear through programming the arduino. Yes! I've got a new data-minion!

Anyhow, expecting great things from Joey. Can't wait to see what he comes up with for the next Makefair.

Using Database Mirror from ASP.NET

2011-06-10T11:13:00.000-07:00

I'm still trying to wrap my head around using Mirroring from ASP.NET. I figured out the Failover Partner part, but the fail over wasn't happening in ASP.NET. Turns out I had left out Pooling=False part, which is key to immediate failover.

FROM:
add connectionstring="Data Source=SQL2010A;Initial Catalog=School;Integrated Security=True" name="SchoolConnectionString" providername="System.Data.SqlClient"

TO:
add connectionstring="Data Source=SQL2010A; Failover Partner=SQL2010B; Catalog=School; Integrated Security=True; Pooling=False" name="SchoolConnectionString" providername="System.Data.SqlClient"

Microsoft states that we should also implement retry logic in our ASP.NET apps, bBut I'm just trying to see what happens with basic apps now.

SQL Mirroring Checklist grows out of Database "x" is not configured for database mirroring. (Microsoft SQL Server, Error: 1416)

2011-06-10T10:25:00.000-07:00

So, I'm still doing some fail over testing,
And I need a new mirror database to play with.
It's been a couple of months since I last worked on this,
How hard can it be?

Well, I made the backup on A,
Restored it to B,
Configured Mirror Security on A,
And tried to start the mirror.
This is what I got back:

Database "x" is not configured for database mirroring. (Microsoft SQL Server, Error: 1416)

What?
What did I do wrong?

Duh!
I left the restore options at defaults,
"Leave the database ready to use..." is the default,
For Mirroring it needs to be set to "Leave the database non-operational..."

Ok, fixed that.
Try mirroring again...

The remote copy of the database "x" has not been rolled forward to a point in time that is encompassed in the local copy of the database log. (Microsoft SQL Server, Error: 1412)

Jumping catfish,
What did I do now?
Duh 2: Didn't restore the log file as well.
Restore a full backup AND transaction log!

What I need is a checklist so I won't have to learn this over again.
And here it is:

Checklist for Mirroring:
1) Take a full backup of the source database.
2) Add a Transaction Log backup to that full backup.
3) Copy the Backup File to Server B.
4) Restore Full and Log to new datase on B.
Be sure to select Options : Check Leave the database non-operational....
5) Back to Server A, Select database, Tasks: Mirror
6) Configure security
7) Start Mirroring

Enabling Ad Hoc Distributed Queries

2011-06-08T10:17:00.000-07:00

One of the developers had a problem today trying to use the OpenRowset command:

Msg 15281, Level 16, State 1, Line 1

SQL Server blocked access to STATEMENT 'OpenRowset/OpenDatasource' of component 'Ad Hoc Distributed Queries' because this component is turned off as part of the security configuration for this server. A system administrator can enable the use of 'Ad Hoc Distributed Queries' by using sp_configure. For more information about enabling 'Ad Hoc Distributed Queries', see "Surface Area Configuration" in SQL Server Books Online.

Of course I've seen this problem in the past, and doesn't even require a server restart to fix. The solution is documented well at http://msdn.microsoft.com/en-us/library/ms187569.aspx, but the script presented has a minor flaw. You are expected to run the script step by step, which, if properly written would not be necessary. So, my contribution to the community is a slightly enhanced script to fix this problem:

USE MASTER
GO
sp_configure 'show advanced options', 1;
GO
RECONFIGURE;
GO
sp_configure 'Ad Hoc Distributed Queries', 1;
GO
RECONFIGURE;
GO

Copy, paste, and run in one shot.

Odd error on webserver after install Dot Net 4.0: The resource cannot be found... eurl.axd/375f4712b7a7e74fae526bfc92711351/

2011-05-17T15:18:00.000-07:00

I recently upgraded our development machines to ASP.NET 4.
We still have a pretty decent investment in ASP.NET 2 Code,
And are just starting to make the jump to Dev Studio 2010.

Something odd on my Dev server though.
After installing Dot Net 4.0 the 2.0 apps,
Especially ones using ASP.NET menus,
Looked a bit odd.
Duh! The server defaulted all WebApps under the default site to 4.0.
Easy enough, just set them to 2.0 in IIS.
Then restart IIS as reccomended.

However, even after doing this,
I attempted to browse to (generic) http://server/app
The full text of the error, and I'm pasting directly so it's searchable:

Server Error in '/app' Application.

--------------------------------------------------------------------------------

The resource cannot be found.

Description: HTTP 404. The resource you are looking for (or one of its dependencies) could have been removed, had its name changed, or is temporarily unavailable. Please review the following URL and make sure that it is spelled correctly.

Requested URL: /app/eurl.axd/375f4712b7a7e74fae526bfc92711351/
--------------------------------------------------------------------------------
Version Information: Microsoft .NET Framework Version:2.0.50727.3620; ASP.NET Version:2.0.50727.3618

Hmm, that's a bit odd.
Default URLs aren't being recognized,
And the Default Document is specified correctly for the application.

Problem appears to come from ASP.NET 4.0:
From: http://connect.microsoft.com/VisualStudio/feedback/details/559174/eurl-axd-httpexception
For a 32 bit server:
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\ASP.NET\4.0.30319.0EnableExtensionlessUrls = (DWORD) 0
For a 64 bit server:
HKEY_LOCAL_MACHINE\Software\Wow6432Node\Microsoft\ASP.NET\v4.0.30319.0\EnableExtensionlessUrls = (DWORD) 0These keys don't exist naturally, you'll have to add them.
Then be sure to restart IIS.

I also found that I could simulate the error by setting the DWORD to 1 and restarting IIS.

Progress on Reporting Services Failover (SQL 2008 R2 SSRS Failover)

2011-05-13T10:00:00.000-07:00

I've been reading a lot of data on SQL Reporting Services 2008 R2 Failover, and wanted to report my progress.

First, I have two SQL 2008 R2 servers in Active / Passive mode (SQL2010A, SQL2010B), with database mirroring active on databases ReportServer and ReporServerTempDB. This is the high availability mode with witness, and I'm seeing about 150ms of overhead. The DR Server, SQL2010B is 2 miles away on fiber.

To configure, I set up reporting services on SQL2010A, which was populated with data from the SQL2005 Reporting Services server. I restored the encryption key from 2005 and everything came up.

I then forced a failover to SQL2010B, and again, restored the encryption key. And confirmed that reporting services is working on B. I then forced a fail back to A.

Reporting services has been left runing on both A and B. Attempting to use reporting services on B will now result in failure, but B will automatically serve requests if the databases is failed over. There is no need to restart reporting services.

I have a CNAME "Reports" pointed at SQL2010A. This simplifies deployement, and allows manual redirection of Reporting services. Navigating to REPORTS reveals reports from SQL2010A. Forcing a failover and updating the CNAME, we're now able to repoint to SQL2010B, and reports continue to be served. Users can still hit http://reports/ after the failover, and in fact, if browser the reports server, the next click transfers to DR.

Next, I've been able to generate a dot net based SQL client. By connecting to both servers A & B as primary and alternate, I can issue a SQL command "SELECT @@SERERNAME" to determine which server actually responded. A simple way to tell which server is active. Reports SQL2010A before failover, and SQL2010B after failover.

Following that up, I've been able to create PowerShell scripts to update the CNAME to either A or B. I'm now working on combining that with the dot net code for an automated fail over solution.

What's left to do:

1) Integrate powershell scripts into dot net failover code.
2) Move subscriptions from Server A to Server B.
3) Clean up subscriptions following a failover.

I've noticed that SSRS automatically rebuilt jobs on the new 2008 R2 server when I migraited from 2005. So there must a routine to do this already. Hopefully there is a procedure to clean up jobs as well.

UAC on Windows Server - A Wicked Bad Idea!

2011-05-11T13:22:00.000-07:00

I'm still supporting one instance of SQL2008 in my enterprise. At the time is was built, we used the current release of Windows Server, which was Windows 2008. Just to be certain, none of this is R2 stuff, just plain Jane 2008.

I needed a development partner for the production machine, so I buillt up a Windows/SQL 2008 in virtual. I remember when I orginally installed SQL 2008 it failed. Well, we tried again in virtual, and it failed again. Odd, I'm a local admin on the box, I installed as Admin, what could the problem be.

My arch nemesis, UAC, just reared its ugly head again. Yes, a fantastic idea on workstations. Kudos to you Microsoft. But it's a really really bad idea on Server. Note to self, check UAC setting before installing. Always make sure it's turned off.

With UAC on, SQL Server 2008 install fails. With it turned off, does not fail. Simple as that.

SQL 2008 R2 Reporting Service Print Client Package

2011-05-10T15:53:00.000-07:00

We have had some serious problems in the enterprise trying to distribute Sql Reporting Services Print Client, or rsClientPrint. We've never quite been able to get the package to push to clients successfully, maybe in part, due to our very restrictive environment. In our environment, users are allowed to run ActiveX controls, but not install them. So, this posting is how we got around the ActiveX download, and directly installed to the workstation.

First, you can find all the install packages on your Sql Server at:
\Program Files\Microsoft SQL Server\MSRS10_50.MSSQLSERVER\Reporting Services\ReportServer\bin

You should find vsersons of rsClientPrint in 3 different flavors.
As of SQL 2008 R2 CU7 these are dated 4-9-11.
RSClientPrint-ia64.cab
RSClientPrint-x64.cab
RSClientPrint-x86.cab

If you expand the x86 cab file you will find:
RSClientPrint-x86-inf
rsclientprint.dll
rsclientprint_1028.rll
rsclientprint_1031.rll
rsclientprint_1033.rll
rsclientprint_1036.rll
rsclientprint_1040.rll
rsclientprint_1041.rll
rsclientprint_1042.rll
rsclientprint_1046.rll
rsclientprint_1049.rll
rsclientprint_2052.rll
rsclientprint_3082.rll

I copied everything but the inf file to a folder on the nas.
\nas2\software\rsclientprint

I then created a simple batch file to install the package (Setup.Bat)
@echo off

echo RSClientPrint Installation:

echo .

echo Copy Main DLL...

xcopy \\nas2\software\rsclientprint\*.DLL C:\Windows\System32 /R /Y

echo .

echo Copy Suppport Files...

xcopy \\nas2\software\rsclientprint\*.RLL C:\Windows\System32 /R /Y

echo .

echo Register Service...

c:\windows\system32\regsvr32.exe /s c:\windows\system32\rsclientprint.dll

echo .

echo Done!

pause

Now, running setup.bat from this folder installs the print client to the current machine.

We were able to use our packaging software to distribute this to all client pcs.

DarkSpore - Dead Again!

2011-05-09T19:25:00.001-07:00

Ok, managed to kill DarkSpore again. Again, EA... Some simple programming would have solved this problem. My computer shut down in mid patch download. I restarted, and the patch continued to download. Great I thought, error recovered. Not so fast. Got an error 6 when the download finished and the system was trying to unzip. EA says to restart Darkspore and try again. Again error 6.

So, where are the bad error files at? Turns out it stores them in Program Files\EA\DarkSpore\DarkSporeBin\Patches. Delete the offending file noted in the dialog, and the system will redownload on next start.

Ad Hoc Querying Oracle 11g 64 bit from SQL 2008 R2 64 bit

2011-05-09T11:27:00.000-07:00

Wanted to query some oracle data from a script today, without first creating a linked server. Worked out the syntax and bumped into some security issues. Thought I'd document so I don't have to look this up again.

First, to turn on distributed ad-hoc queries:

exec sp_configure 'show advanced options', 1;

RECONFIGURE;
exec sp_configure 'Ad Hoc Distributed Queries', 1;
RECONFIGURE;
GO

Next, to directly query from Oracle 11g 64 bit from SQL 2008 R2 64 bit without a Linked Server:

Select * from OpenRowSet('OraOLEDB.Oracle','ORA.WORLD';'user';'password','SELECT * FROM schema.view')

See my previous post for setting up Oracle 64 Bit Client on SQL 2008 R2 64 bit.

Darkspore already dead....

2011-05-08T22:01:00.000-07:00

Picked up DarkSpore over the weekend. The sequel to Spore I had some better hopes for it. It's turned out to be more like a cross between poke-e-mon and spore at this point. Very dissapointing that the program failed to start today. Reinstall did not work. Turns out that the local data cache was corrupt. Under Windows 7 thats \users\username\AppData\romaining\DarkSporeData. Just delete the folder and restart the game. Come on EA. You can have a much better error recovery procedure than that. Set a dirty start flag. If the game is restarted with the dirty flag set, then offer to fix. Gradeschool programming!

SQL 2008 R2 Bringing a "Restoring..." database back online...

2011-05-06T15:04:00.000-07:00

I've been working with Disaster Recovery scenarios in SQL 2008 R2:

We all know that the mirror database starts In Recovery, and moves to Restoring while being participating in the mirror operation. But what happens when the mirror breaks? In that case, the database stays in Restoring mode. If viewing from your workstation, you'll just still see Restoring. If viewing from the server, you'll at least see that the mirror has been removed. In this state, the database does not seem to be of much use. You can't attach or read from it.

In my scenario, I want to the bring the Restoring / Broken Mirror database back Online. Let's assume that the principal failed and the Mirror did not have a chance to take over, for whatever reason. How do we get the Restoring database back online?

Pretty simple really. Issue the following commands:

Use Master

Restore Database X (Where X is the name of the db)

Realize that this creates a separate fork point in recovery from the original principal database. But if you need that database, and need it online now, this is the way.

SQL 2008 R2 SP 1 Coming!

2011-05-02T10:46:00.000-07:00

Read in a blog today that SQL 2008 R2 SP1 just hit CTP.
Looks like SP1 also contains a few enhancements.
Problem is, it's only supporting CU1-CU6, and I just finished updating to CU7!
Looks like I have to wait for the next CU to come out after the SP to get SP1.

SQL 2008 R2 64 Talking to Oracle 11g 64 bit

2011-04-29T15:49:00.000-07:00

Back to this subject again.

I thought I worked this out before, but it looks like I need to a do a little 1984 editing to one of my previous postings.

I tried to run the following on SQL 2008 R2 64 with an Oracle 11g 64 bit client installed. (Of course details have been moved to protect the innocent):

exec sp_addlinkedserver 'ORATEST', 'Oracle', 'MSDAORA', 'ORACLE.WORLD'

exec sp_addlinkedsrvlogin 'ORATEST', false, NULL, 'user', 'pass'

select * from openquery(ORATEST,'SELECT * FROM USER_OBJECTS')

exec sp_dropserver 'ORATEST', 'droplogins'

Well, I got a big whopping failure with:

Msg 7308, Level 16, State 1, Line 1
OLE DB provider 'MSDAORA' cannot be used for distributed queries because the provider is configured to run in single-threaded apartment mode.

What gives?

Der, MSDAORA is a 32 bit provider.

Dummy, you just installed 64 bit 'OraOLEDB.Oracle' provider and you're not using it!

So this really works:

exec sp_addlinkedserver 'ORATEST', 'Oracle', 'OraOLEDB.Oracle', 'ORACLE.WORLD'

exec sp_addlinkedsrvlogin 'ORATEST', false, NULL, 'user', 'pass'

select * from openquery(ORATEST,'SELECT * FROM USER_OBJECTS')

exec sp_dropserver 'ORATEST', 'droplogins'

SQL 2008 R2 Database Mirroring Scenario

2011-04-27T10:57:00.000-07:00

I've decided to go with Database Mirror in SQL 2008 R2 for our DR Scenario. The plan is to migrate database forward from SQL 2005 to SQL 2008 R2, Starting with the Upsized Access projects and the ASP.NET sites.

Both SQL 2008 R2 servers are matched equipment, 16 processor monsters. SQL2010A lives in our production center. SQL2010B lives in our DR center 2 miles away, but linked by fiber.

Current scenario considerations are:

Need to provide automatic failover to standalone DR site if primary data center fails.
Need to maintain connection at primary data center if comm line to DR center fails.

Configuration:
Primary Server: A
Mirror Server: B
Install Witness W on Mirror Server B as a second instance.
So, for simplicity, servers are A, and B/W.

Fail Over Scenarious:
O=Online, X=Offline

Case	A	B	W	Master
1	O	O	O	A
2	O	X	O	A
3	O	X	X	A
4	O	O	X	A
5	X	O	O	B

Scenario 3 was my chief concern,
With a total loss of communication with the DR Center,
Main server operation continues.

Technically failover is only permitted,
If B and W are reachable and A is offline.

Looking for SQL 2008 R2 64 Bit patches (CU 7)

2011-04-26T12:17:00.000-07:00

I've been having some problems with Reporting Services on SQL 2008 R2. Seems there's litterally been a couple hundred patches to SQL 2008 R2, and still no Service Pack. So it's time to start looking very carefully at the CUs. There's been 7 CUs to date. 7! That's for a product released 9 months ago. That's almost a new set of patches every month.

Well, I started going nuts looking for the 64 bit versions of the patches. I'm seeing 32 bit versions easily enough, but couldn't seem to find any 64 bit versions. For the record, the 32 bit patches will not work against the 64 bit engine. Makes sense?

Well, Microsoft out thunk me on this one. On the patch download page, there's a "Show hotfixes for all platforms and languages" link. As my desktop is 32 bit, and my server 64 bit, only the 32 bit patches were shown. (For safety, we only like downloading to workstation, then move to server.)

As another suggestion, always save all patches, on server, and in archive. There's no telling when Microsoft may pull a CU, and not allow future download. Sometimes, we need to build a server back to exact version to be able to restore master db. So hang onto those CUs!

SQL 2008 R2 CU 7
Build 10.50.1777.0
http://support.microsoft.com/kb/2507770

Missing Disk Space in Windows Server 2008 R2 64 + Sql Server 2008 R2 64: Revisited

2011-04-12T13:59:00.000-07:00

Wanted to update on what we finally found out this "bug." Intentional or not, it still feels like a bug when a system administrator is locked out of any part of a machine. Well, you've probably guessed, it's UAC, User Access Control, in Server 2008 R2 causing our problems. It's easy enough to disable, and we can see all the files on the server again. I understand the reasoning behind UAC, but not sure that I can see its place on Servers.

Missing Disk Space in Windows Server 2008 R2 64 + Sql Server 2008 R2 64

2011-04-11T14:03:00.000-07:00

I've been chasing down a problem where a huge chunk of disk space went missing on a Windows Server 2008 R2 64 box with Sql Server 2008 R2 64 installed. Logged in as net admin, Windows reports 27gb of disk usage of a 200gb drive. Problem is, only 20gb free. Where's the missing data?

We discovered that by logging into the server through a network file share file://server/c$, we were able to see all 117gb of disk usage. Yah, that's still a bit off from 200, (paging files and such) but still, there's the missing 90+gb. We did a search on the share for large files, and a few MDF files popped up. Big ones.

SQL Server had written replicated databases to its C: drive, and then masked them from admin view. That's just plain wrong. Cruising the data folder, and taking permission of it, allowed the files to be counted. Now that is just wrong. The Domain Account servicing SQL Server is in the network group DB_ADMINS. DB_ADMINS are in the Local Admin Group. We should have the same god like perms on this server.

So, if your SQL 2008 Server is mysteriously missing disk space, check your db folders.

Mounting AVR ports to Front or Back of a Board

2011-04-09T17:07:00.000-07:00

I've been working on designs for AVR connections to printed circuit boards, and have been getting really confused in the process. You see, I'm working with single sided boards. So, somtimes the AVR plug will mount to the front of the board, and sometimes, to the back of the board.

I know that sound a bit confusing, but consider the following design for a multi-chip adapter for an AVR MK-II type USB burner. The circuit layer is on the bottom of the board, the parts are connecting to the front, and everything is being soldered on the bottom.

In this case, the following layout describes which signals to expect when you properly connect the USB Plug from the burner:

The problem occurs, and it's probably due to my own stupidity, when you try to reverse the situation. That is, the circuit represents the top side of the board, and you are trying to mount the ISP port on the bottom of the board. In my case, this is for an AVR ISP Breakout board.

If you stop and think about this for a minute, it's the exact mirror of the previous situation. Spacially, we've rotated the plug around 180 degrees. And yes, from trial and error, it really is what works.

So why am I documenting this? Well it's not the first time I've run into the this problem, nor will it most likely be the last. Future self, please remember to reverse the port if mounting ISP port on the back...

If you want to see more about any of these boards,
including schematics and board layouts,
please see SMTboards.Com

Problems installing IE9

2011-04-09T16:23:00.000-07:00

Looks like IE 9 was released to the public on 4/5/11 through windows update. Ok, I'm game, if it's released then it should work, right? Or so I thought...

I've just had the joy of uninstalling IE9 twice. My Windows 7 Professional equipped laptop is otherwise fully up to date.

My first impression is that IE9 is faster than the Wako Kid. Remember the movie Blazing Saddles? Gene Wilder played the Wako Kid, a whiskey guzzling gunslinger that sobers up, and moves so fast that he never even moves. That's sort of like IE9. Won't start. Won't allow programs that use it to start either. Pretty much a dead system without a browser. However, unlike the Wako Kid, IE9 won't sober up.

I checked the web, from another machine, and downloaded install method two. That's an IE9 executable with "everything required to install IE9." This lovely gem started, but stated that a new version was installed on the computer and it was quitting.

But, here's kudos to Micorosft, IE9 does cleanly uninstall if you know how.
Control Panel | Programs and Features | View Installed Updates

Scroll down to Microsoft Windows at the bottom.

Most recent update should be Internet Explorer 9.
Uninstall it. A reboot is required with two stage patch removal.
When you're back up again, you're back to IE 8.
Thank you uninstall.

And just so I don't get the hate mail, I did uninstall, reboot, then try that IE9 executable again. It did install, reboot, but again, brain dead on arrival. Abby, abby something... Abi-Normal!

But here's where I'm really pissed about this whole thing. And here's the lesson learned. If you're going to update a critical program like IE, make sure it works after the update. Why did I, the user, have to discover that IE9 wasn't working? Shouldn't the installer have detected this and taken corrective actions? This is the 21st century after all.

Fixing a locked hosts file.

2011-03-20T16:18:00.001-07:00

I have been struggling with an infected windows 7 laptop for some time. Cleaned virus Perfect defender, piece of junk, but it left a couple hundred bogus entries in the hosts file. Then locked it out from change.

But this is windows right? I'm the admin. How can I be locked out of any file?

Jerk faces set read only, system, hidden on the file. No security tab appears.

The trick is to boot into safe mode. Click advanced. Ownership. And take ownership of the file. Close dialog to save changes.

Reedit the file, now set full perms for authenticated users. Close dialog to save changes.

Now delete the damn file and reboot. Windows recreates the file if needed.

Published with Blogger-droid v1.6.7

Why doesn't news update in Visual Studio 2005?

2011-03-18T12:15:00.000-07:00

Ok, call me old fashioned, but I'm still using Visual Studio 2005. Frankly, 2008 was too buggy. And I'm holding out on my Windows 7 box to install 2010. So, I'm keeping Visual Studio 2005 around for compatability reasons. (Yes, Yes, I know 2010 is supposed to be compatible, but 2008 did some horrible things to my ASP.NET code.)

One thing that has been bugging me is that the news on the start page never seems to update. Duh, I thought, I bet Microsoft deprecated a URL and never updated Visual Studio. I was right. The RSS feel for SQL Server changed, and here's how to fix it.

From Visual Studio 2005
Go to Tools | Options | Environment | Startup
Observe the Start Page news channel url.

Navigate to your MSDN page of choice in an new browser window.
For me that's MSDN.Microsoft.Com/SQL
Notice the RSS Feed symbol on the page.
Right click and copy URL.

Back to Visual Studio,
Paste that URL into the start page new chanel textbox.

Wala, you have current news...

What I learned on day 1 of moving from Rock Band 3 Mustang to Squier Guitar...

2011-03-10T23:24:00.000-08:00

I read a lot of false information about the Squier Guitar for Rock Band 3. Here's what I found on day 1.

0) In the US at least, the only place to get the guitar is from Best Buy. They originally screwed up and shipped the first batch directly to stores bypassing pre orders. No apology given. Horrible company to deal with. Places a continuous hold on your credit card while waiting on pre order. Will never pre order from Worst Buy again.

1) The Squier does not work in game without the MIDI controller. If you can't find one, don't bother getting the guitar. Amazon seems to have them in stock right now.

2) You do not need to buy a MIDI cable, as one comes with the guitar. Though a really short one <6" would be really nice. I now have a large cable loop at the bottom of the guitar.

3) You are not wireless. Yup, you cable from the Guitar to the Midi Controller, to the XBOX. Kinda lame, should have been wireless.

4) You cannot use the Rock Band 2 Overdrive pedal, as there's no place to plug it in. Though there is an overdrive button conveniently located on the wrong side of the guitar. I'd like to see someone try to hit that while playing. I'm going to work on modifying the MIDI controller to accept the Overdrive pedal.

5) Overdrive is built into the Guitar, ala Tilt Up.

6) We already knew this, but it goes without stating, there is no Whammy bar on Squier, hence no way to wa wa notes or boost your overdrive points.

7) The MIDI controller mounts directly to the bottom of the guitar. Basically you loosen the bottom strap nut, slide on the midi adapter, and tighten to clamp the controller in place.

8) The Xbox buttons on the guitar are non-functional in game, you must use the MIDI controller buttons. Apparently this is something to do with an XBOX rule about buttons must be on controller. Big notice in the box with the Guitar. This seems really dumb. The buttons do function to control certain synthesizer outputs of the Guitar in MIDI mode, not sure if this does anything in game.

9) There's a mute-bridge on the guitar, and it's suggested that you engage this when in game. This prevents false readings of overstrums. This also prevents you from plugging in an amp while you play. Lesson learned, the guys playing the guitar in game at the demos were really really good. For schmucks like me, it's probably mute-bridge to gett better scores.

10) I learned to pluck outwards on the Mustang, this does not work on the Squier. You must pluck downwards for the note to register.

11) Er, duh, there's an on/off switch on the back of the guitar. You have to manually turn the thing on and off to play in game. It's not required to be on if using an amp. But I'm sure I'm going to forget and leave the thing on. 3 AA batteries, supplied with the guitar.

12) You've gotta learn how to tune a guitar. Even Androids Don't Go Behind Elephants. E A D G B E. Get a digital tuner, you'll be really happy you did. They're cheap, about $12 on Amazon.

13) Ok, and don't rip me for this, my fingers hurt. This is a real guitar and you have to push down on the strings. My programmer fingers are a bit sore, even after playing with the Mustang for several months.

14) There are two hex keys in the box, these are for adjusting the bow of the neck, and for adjusting the height of the strings. Find a safe place to keep these.

15) Its highly suggested that you use a pick. I've been playing with my tumb, so I could feel the strings better. Guess it's time to start learning to pick.

16) Er, Can't say that the guitar really sounds good coming through my crappy little 9v battery powered amp. It's because there isn't much there begin with on Beginner Pro Guitar. Time to ramp it up a bit. Still, you can hear that you ARE playing some kind of music, just doesn't sound as good as the Rock Band track.

Woot! Woot! Rock Band 3 Squier Guitar Arrives!

2011-03-10T18:50:00.001-08:00

The Squier finally arrived! Yes the MIDI controller is absolutely required. Yes, a MIDI cable is needed. BUT, it does come with the guitar. The MIDI controller does mount directly on the guitar. Sadly, Microsoft did not allow enabling of the game controller buttons directly on the guitar. This means you have to use the buttons on the MIDI controller. But because its mounted at the bottom of the guitar, the direction of the dpad is odd. Microsoft please amend this really dumb policy.

Published with Blogger-droid v1.6.7

Rock Band 3 Squier Update

2011-03-09T11:47:00.001-08:00

UPS has the package out for delivery from sparks nv. that's about 50 miles south of Vegas. gonna be a long drive home for the UPS driver

Published with Blogger-droid v1.6.7

Bestbuy still can't get Rock Band 3 Squier Guitar right

2011-03-08T16:17:00.001-08:00

Received email notification yesterday at 1:16 PST that Best Buy has shipped the Squier Guitar from my preorder. Great! Maybe they finally figured out what preorder means.

But wait. 27 Hours later and UPS still says it hasn't received the package. BB customer service was only able to read the status screen verbatim.

Rock Band? You put your faith in the wrong company to exclusive rights on this product. I would have much rather gone through Amazon.

Published with Blogger-droid v1.6.7

Besr Buy side stepped pre orers

2011-03-03T14:19:00.001-08:00

According to the Rock Band forum, Best Buy did side step all the pre orders and shipped Rock Band 3 Squire guitars directly to stores. No apology. Very nice Best Buy. You officially suck.

Published with Blogger-droid v1.6.7

Rock Band 3 Squier Guitar Backordered

2011-03-01T23:34:00.001-08:00

Well, I blogged previously about harmonix and Rock Band 3. Awesome game, but buggy at launch LAST OCTOBER. Well, its 5 months later and still no squier real guitar. ~~Its been available for months in England. But not the US *.~~ Best Buy has been given exclusive marketing rights in the US and boy do they suck. Constantly purchasing on the card even though they are not shiping. No word if they even got stock. Just your pre order is back ordered. I‘ve never ever seen that happen before.

But the fun does not end there. Harmonic is not producing enough midi controllers. Last word was something like 5000 may have hit the US.

What the Frack is the problem? Harmonix you have a captive and very dissatisfied customer base. What are you going to do about it?

* Redacted. False rumour. No squier guitars to be had in the UK either. So where the frack are they?

Published with Blogger-droid v1.6.7

Formula for cheap board etching

2011-02-23T22:18:00.001-08:00

Use eaglecad. Print on laser to transparancy. Only layers needed are top, pads, and vias.

Lay out on glass master slide so that readable side faces glass to protect toner. Sandwich board between two glass sheets and clamp.

Expose 12 minutes. Use red light when handling outside exposure if possible.

10 cups water to 1 cup positive developer. Do not over develop. Room temperature is just fine.

Add 4 cups muratic acid to 2 quarts hydrogen peroxide.

Use quality nitrile gloves as this stuff really burns and leaves nasty yellow strains.

Published with Blogger-droid v1.6.7

What ever happened to Voyager 1?

2011-02-22T10:30:00.001-08:00

Voyager 1 is the fastest moving thing ever created by man. its travelling at 85,000 mph. So far traveled 14 billion miles. Even at that rate its only gone .001 light years away.

Not sure if the Vger shock mind melded with was one or two... But that looks a long way off...

Published with Blogger-droid v1.6.7

Manual Removal of Oracle Client 11g

2011-02-14T19:50:00.001-08:00

In preparing the previous posting,
I've had to do a lot of installing and uninstalling of Oracle Client.
The Oracle installer is by far the worst installer I've ever seen for windows. And a 1gb installation for any driver is insane.

1) Firstly, I always install to C:\app
Oracle would like to isntall to C:\app\USERNAME
Where username is your login id.
Anyhow, to make the install more generic, I force it to C:\app

2) Once installed, this thing spiders out like a virus.
The uninstaller is just as crappy.
So, manual install is the best way to get rid of this client.
Problem is, there's a lot of references in the registry,
Many of which prevent you from removing the client.
Here's my uninstall method:

Using Regedit
1) Remove HKLM\Software\Oracle
2) Remove HKLM\SYSTEM\currentControlSet\Services\Ora*
May be 1 or 2 Recognizable Oracle services there.
3) Here's where it gets a bit tricky.
HKLM\Software\Microsoft\MSDTC\MTxOCI
Add an x before each string for OracleOciLib, OracleSqlLib, and OracleXaLib
Otherwise, these files, if they exist, will still be in use next time you reboot,
And unable to be deleted.

4) Start | Right Click My Computer | Properties | Advanced Settings | Environmental Variables
Edit the path varialbe and remove Oracle references.
5) Restart

6) YOu can now delete the program files which were previously open.
Delete C:\Program Files\Oracle
Delete C:\App
7) Restart just to be sure.

8) Remember, if you reinstall Oracle,
You'll need to correct the keys in HKLM\Software\Microsoft\MSDTC\MTxOCI
See my posting yesterday for more info.

Published with Blogger-droid v1.6.7

Connecting SQL Server 64 bit to Oracle

2011-02-10T14:01:00.000-08:00

I've had a ton of difficulty connecting from SQL Server 2008 & SQL Server 2008 R2 to Oracle.
Of course the majority of the project has been the 64 bit support.
I've finally gotten it to work, cleanly and simply...

1) Install & Test SQL 2008 R2 x64

2) Install & Test Oracle 11g 64 Bit Client, Admin Install.
Be sure to update and test the TNSNAMES.ORA file.

3) Edit the following keys:
HKLM\Software\Microsoft\MSDTC\MTxOCI
OracleOciLib = oci.dll
OracleSqlLib = orasql11.dll
OracleXaLib = oraclient11.dll
& Reboot.

4) Under SSMS:
Server Objects | Linked Servers | Providers | OraOLEDB.Oracle
Right Click | Properties |Check: Allow inprocess

Now, in a script window:

use master

exec sp_addlinkedserver 'LINKED_SERVER_HANDLE', 'Oracle', 'ORAOleDB.Oracle', 'TNS_NAME_ENTRY'
Where: LINKED_SERVER_HANDLE is the name to reference this entry by in future.
TNS_NAME_ENTRY is the Oracle TNS Name Entry of the Server to Connect to

exec sp_addlinkedsrvlogin 'LINKED_SERVER_HANDLE', false, NULL, 'USERID', 'PASSWORD'
Where: LINKED_SERVER_HANDLE is that same reference.
USERID is the Oracle User ID.
PASSWORD is the Oracle Password.
Note: This gants all users capable of logging onto SQL permission to use this connection.

select *
from openquery(FIN,'SELECT * FROM SCHEMA.ORACLE_TABLE_OR_VIEW')
Where: ORACLE_TABLE_OR_VIEW is a valid view in available the log in credentials used.

ie:
exec sp_addlinkedserver 'FIN', 'Oracle', 'OraOLEDB.Oracle', 'FIN'
exec sp_addlinkedsrvlogin 'FIN', false, NULL, 'finuser', 'finpass'
select * from openquery(FIN,'SELECT * FROM APPS.EMPLOYEE_V')

Information partially gleemed from:
http://serverfault.com/questions/175257/how-to-install-oracle-ole-drivers-for-use-in-sql-server-2008-r2-x64

Adding Report Manager to an Exsting SQL 2005 Express install: Error loading assembly PublicKeyToken=b03f5f7f11d50a3a'

2011-01-14T10:19:00.000-08:00

Received the following error while trying to install Report Manger on Sql 2005 Express (SP1):

Message: Could not load file or assembly 'System.Web, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a' or one of its dependencies. Access is denied.

Problem was read/write permissions on the report manager folder:

Assign & Propagate Full Permission to Network Services on the following folder:
C:\Program Files\Microsoft SQL Server\MSSQL.2\Reporting Services\ReportManager

Copying Sql Reporting Services 2005 to Test Server

2011-01-14T08:46:00.000-08:00

Working on on an update to one of my systems, I decided to copy Sql Reporting Services 2005 from my production server to the test server. I just didn't feel like recreating that mass of reports, and heck, it's just a database right?

Steps to copy Reporting Services:
1) Back up ReportServer and RepportServerTempDB on Production.
2) Back up the Reporting Services Encryption Key using Reporting Services Configuration tool on Production.
3) Copy these files (backups + key backup) to your test server.
4) Stop Reporting Services by using Sql Server Configuration Manager
5) Delete or rename existing ReportServer and RepportServerTempDB
6) Restore ReportServer and RepportServerTempDB

7) Restart Reporting Services by using Sql Server Configuration Manager

8) Restore Encryption Key using Reporting Services Configuration tool.

Well, the only problem with this is that all the scheduled jobs also moved over and started executing from the test server as well. Well, that's not really so bad. It proves the system will rebuild scheduled jobs if it needs to, but I really don't need the users getting 2 copies of their reports. I really didn't want to disable the mail server on the test system, what was needed was a simple way to delete all jobs.

REVISESD: Delete all jobs on Test Server:
(I originally missed ReportSchedule, thanks SqlDisco!)

Use ReportServer
DELETE FROM ReportSchedule
GO
DELETE FROM Subscriptions
GO
DELETE FROM Schedule
GO
SQL automagically deletes all the scheduled jobs in Sql Server Agent as well!

Kudos and thanks to multiple sources,
Books Online
And http://www.sqldisco.com/?p=44

Boot Linux from USB Stick

2010-12-07T18:35:00.000-08:00

A neighbor of mine just purchased a large number of PC's at auction. Only problem, they all come without hard drives. What's the cheapest way of getting a working operating system on these computers?

1) Boot Ubuntu 10.10
2) Insert a 4g stick
3) Run System | Administration | Startup Disk Creator
4) Reboot with stick still in drive, select boot from USB

5) Once booted, at the Install/Trial screen, just hit X.
6) Run Applications | Accessories | Terminal | sudo apt-get purge ubiquity
7) Delete Install 10.10 Icon from the desktop
8) Shutdown

You know have a working Linux Boot Stick without Trial Nag!

Thank you to:
http://ubuntu-virginia.ubuntuforums.org/showthread.php?t=1598143

Getting Linux to run on sub $100 WM8505 Netbook

2010-11-19T13:35:00.000-08:00

I bought a couple Smart Book netbooks for just around $100 each. They weem to be dropping as low as $75 if you can catch them right. The only real problems so far is that the Adobe Flash is out of date, and some websites are refusing the browser in preference of IE 8. These are pretty much show stoppers in my opinion, so I've been working on getting Linux to run on it.

My first break was this thread:
http://ubuntuforums.org/showthread.php?t=1349626&page=101

Which led me here:
http://bur.st/~abrasive/wm8505_linux/

I've gotten that core to load by following the instructions in the README.
I'm using a 2gb sd card, which mounts upsidedown in the left side of the system.
Problems I've experienced so far:

1) Now that I've got command line Linux working, now what?
I copied all the scripts from abrasive into folder /usr/WM8505 using another linux machine with the SD card mounted. Renamed all the .txt files to .sh. That should make configuration easier.

2) Instructions say "after you set up your network," how the heck do you that?
Most simply put, you need to create /etc/wpa_supplication.conf
Directions from here:
https://wiki.archlinux.org/index.php/Wpa_supplicant

A) Run the following command to convert your network password to hex:
wpa_passphrase mywireless "secretpassphrase" > /etc/wpa_supplicant.conf

B) Edit the resuling file to look like this:
vim /etc/wpa_supplicant.conf
network={

ssid="mywireless"
proto=RSN
key_mgmt=WPA_PSK
pairwise=CCMP TKIP
group=CCMP TKIP
psk=7b271c9a7c8a6ac07d12403a1f0792d7d92b5957ff8dfd56481ced43ec6a6515
}

C) Add configuration (untested) for hard wire lan.
vim /etc/wpa_supplicant.conf

network={

key_mgmt=IEEE8021X
eap=MD5
identity="user"
password="password"
eapol_flags=0
}

3) Run the wi-on.sh script in the WM8505 folder.
These are the scripts you copied in earlier.
Should connect to your wifi or hardwire.
Look for errors about reading or parsing /etc/wpa_supplicant.conf
It's trying to tell you what the problems are.
Break and rerun till you get it right.

4) Next, I want to install xwindows,
So, folling the instructions in INSTALL.X11
apt-get install xserver-xorg-video-fbdev xfonts-base xinit

I immediately got an error:
Segmentation Faulty Tree
Being thrown by apt-get.
What the?

Researching, it's corrupted indexes for apt-get:
http://www.deanlee.cn/linux/apt-get-segmentation-faulty-tree/

delete /var/cache/apt/*.bin
apt-get update

Then try the install again.

5) That worked, and I tried copying in the useful xorg.conf from abrasive.
A) Backup the original configuration
cd /etc/X11
cp xorg.conf xorg.conf.bkp
B) copy in the updated version
cp /usr/WM8505/xorg.conf .
C) try to start xserver
xinit

Unfortunately, crashes horribly.

6) I checked the /etc/X11/xorg.conf.bkp file,
Seems to contain the appropriate driver,
So I put the orginal back in place.
cp xorg.conf.bkp xorg.conf

I also installed a few more packages
apt-get install xterm x11-utils xutils xauth

Typining in xinit brings up xwindows,
But not very interesting.

Rockmoninov my fury tail!

2010-11-15T09:16:00.000-08:00

In the 1971 masterpiece Willy Wonka and the Chocolate Factory, Mrs. TV, the Geography Teacher, states that the combination to the musical lock is Rachmaninov. Never trust a Geography Teacher to know her music. The piece is actually Mozart, the first few bars of Marriage of Figaro (Overture).

Ok, it may not have been bugging you very much, but it's been bugging me for about 39 years. I finally played the movie and the piece side by side on my computer last night just to make sure.

Ok, I feel much better now. Mozart finally gets his movie wrong righted. You go Wolfgang! We love you!

Beat rock band 3

2010-11-05T15:41:00.001-07:00

Well, completed rockband 3. nice surprise ending. Good Job harmonix. Please fix your website. its been broken since launch a week ago

Published with Blogger-droid v1.6.5

Triple release on XBOX (Fable3, Fallout3Vegas, Rockband3) awesome, but flawed.

2010-11-03T09:51:00.000-07:00

To anyone reading this, all of these three games are fantastic technological advancements. The studios should be very proud of what they created, but, have also demonstrated a lack of organizational skills in producing a final product. With that said...
I'm really disapointed by the lack of QA in the recent triple release from Microsoft Game Studios of RockBand 3, Fallout 3, and Fable 3. All three games are incomplete and have serious bugs or omissions that fail to fulfil the promises made by the publisher.

RockBand 3:
While an awesome technlogical achievement, the product channel was not ready for this game launch. Phyisical items, the Guitars, and virtual support, the Website, are not available or broken. The detail: The Keyboard was available a month before launch, Bravo MadCatz! But the Mustang Guitar (MadCatz) won't be out for 2 more weeks. And the Squire Real Guitar (Fender) won't be out for 5 months (3/1/11). Ok, that's just poor marketing. Let's add to this that the RockBand website is broken, and has not allowed account linking (game to website) since the launch. Exporting RockBand 2 songs was initially broken, but was fixed within the first 24 hours. Now, for the past 24 hours, the site will not allow logons at all. Who tested this? I'll give RockBand 3 kudos for a successful game on the 360, but negative marks for poor marketing and poorer web support. As this is a web linked game, I'd say 1/2 of the experience is not working.
Fallout 3 New Vegas:
Again, another fantastic game from Bethesda Softworks, but serious game flaws once again. Gamers are reporting save game corruptions, freezes, and several annoying glitches such as phantom loss of companions. Again, this comes down to really poor QA.

Fable 3:
Again, my kudos to LionHead Studios for another amazing chapter. Unfortunately, plagued by the same problems as Fallout 3, save game corruptions and freezes. There are areas in the game that appear to be unreachable, unless you are very lucky to stumble across them. All in all, I know Lionhead spent a lot of time on this code, but it obviously wasn't QAd properly.

I'm going to point all these problems back to Microsoft Game Studios. Microsoft, as publisher of these games, consumers have an expection that the game will be free of bugs. Game stoppers like I've mentioned detract from the value of the games and the platform as a whole. I'm a beta tester, I can handle unstable releases. But this is not a beta. This is production.

Microsoft, what are you going to do fix these problems and prevent it from occurring on the next game release?

Fallout New Vegas

2010-10-24T23:39:00.001-07:00

Taking stone time of projects to play fallout new vegas. just trying to make it home to Henderson is tough. One thing is abbundently clear... I need much bigger weapons! Where's Mothership Zeta sheen you really need it!

Published with Blogger-droid v1.6.3

Depricated in XNA 4.0

2010-10-05T11:36:00.000-07:00

Working on my code for CommBatt again... Trying now to make it run in the new XNA 4.0. Man, did they break a lot of stuff in this version. Once upon a time, if you were going to change a procedure, you at least left the old version around for a rev. Nope. in this version of XNA they just yanked it out. Ok, I can see the simplifications, but guys? Most of this stuff has already survived through 4 revisions. There's going to be a lot of broken code!!!

1) TotalRealTime no longer exists.
    Hoping that TotalGameTime is drop in.
    Yes, I know the difference.

2) GamerPicture is gone.
    You now get a stream object rather than a Texture2D object.
    That's easily converted with:
     //gamerPicture = profile.GamerPicture;
     gamerPicture = Texture2D.FromStream(GraphicsDevice, profile.GetGamerPicture());
     Unfortunately, my subclass did not have access to the GraphicsDevice, after all, it was a subclass.
     So, I need to pass in the GraphicsDevice now... oh well.

3) Constructor for Texture2D no longer takes Texture.None.
    //pixel = new Texture2D(graphicsDevice, 1, 1, 1, TextureUsage.None, SurfaceFormat.Color);
      pixel = new Texture2D(graphicsDevice, 1, 1);

4) Color moved around in the FrameWork.
   // Microsoft.Xna.Framework.Graphics.Color c;
   Microsoft.Xna.Framework.Color c;

5) Storage.TitleLocation broken?
    Oh joy, welcome to IsolatedStorage class.
    Reference: http://machaira.spaces.live.com/blog/cns!C6A0309746469D9E!724.entry
    //String StageIndexPath = StorageContainer.TitleLocation + "\\Content\\" + file;
    //StreamReader streamReader = new StreamReader(fs);
    byte[] data = new byte[TitleContainer.OpenStream(file).Length];
    StreamReader sr = new StreamReader(TitleContainer.OpenStream(file));

6) Very odd, but references to ArrayList are broken?
     //ArrayList vectors;
     List <Vector2>= new List<Vector2>();

Things to be aware of when Moving Sql Server 2008 Location.

2010-10-04T10:05:00.000-07:00

I recently had the joy of moving a SQL Server 2008 instance from D: to C:. We originally installed on D: as C: was only 20g. No bid deal I thought, after all, it was a 600gb raid 5 array with 20g to C: and the rest to D:.

For the record, system files, including program files, belong on C:. This simplifies later resizing of the C: drive should it be necessary. In the case of Sql Server, having only Data on D: would have simplified my life, I could have simply recreated the drive and restored from backup.

Small aside here: Server 2008 will not allow you to resize a System or Boot partition. That's pretty damn worthless in my opinion. Nice that you can resize anything but C:! The work around is boot Ubuntu (10.04 currently), and execute GPARTED. You will be allowed to resized as you need. Windows will do a checkdisk and readjust when it boots.

Anyhow, things to be aware when doing a Sql System Recovery, at least from 2008.
0) You should already have backups, tested backups, of Master, Model, MSDB & all user data.
1) Use Reporting Services Configuration Manager to backup the encryption key in case anything happens.
2) For safety, shut down the instance and make copies of the master, model, msdb, and tempdb off drive.
3) Uninstall the old server, reinstall the new server on C:.
4) Use Sql Server Configuration to disable TCP/IP on the new server (so no one steals your single user)
5) Start in single user mode:
C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\BINN\SqlServr -m

-- Move the master database
RESTORE DATABASE [master] FROM

DISK = N'D:\Restore\master_backup_2010_09_30_162401_6956296.bak' WITH FILE = 1,
MOVE N'modeldev' TO N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\master.mdf',
MOVE N'modellog' TO N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\master.ldf', NOUNLOAD, REPLACE, STATS = 10
GO
-- Forces server restart...

6. Make sure Sql Server Agent is stopped, or you won't be able to import MSDB.

-- Move in MSDB, contains all your scheduled jobs, packages, and maintence tasks.
RESTORE DATABASE [msdb]
FROM DISK = N'D:\Restore\msdb_backup_2010_09_30_162401_7546355.bak' WITH FILE = 1,
MOVE N'MSDBData' TO N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\MSDBData.mdf',
MOVE N'MSDBLog' TO N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\MSDBLog.ldf',
NOUNLOAD, REPLACE, STATS = 10
GO

7. Update linkages for model & tempdb (hard wired in Master)
-- Copy your model & tempdb from the hard file copy to C:
-- Then use the following to update SQL on the new locations.

ALTER DATABASE model
ODIFY FILE ( NAME = 'modeldev' , FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\model.mdf' )
GO

ALTER DATABASE model
MODIFY FILE ( NAME = 'modellog' , FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\modellog.ldf' )
GO

ALTER DATABASE tempdb
MODIFY FILE ( NAME = 'tempdev' , FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\tempdb.mdf' )
GO

ALTER DATABASE tempdb
MODIFY FILE ( NAME = 'templog' , FILENAME = N'C:\Program Files\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\DATA\templog.ldf' )
GO

8. Various fixups....

-- DLL locations hard stored in MSDB, moving location fails backups, this is the fix.
-- Per: http://support.microsoft.com/?kbid=914171
use msdb
go
delete from msdb.dbo.syssubsystems
exec msdb.dbo.sp_verify_subsystems 1
go

-- After moving MSDB location, backup jobs fail when attempting to save history.
-- Purging all job history solved the problem:
-- Original code http://blog.sqlauthority.com/2008/11/11/sql-server-delete-backup-history-cleanup-backup-history/
USE msdb
GO
DECLARE @DaysToKeepHistory DATETIME
SET @DaysToKeepHistory = CONVERT(VARCHAR(10), DATEADD(dd, 0, GETDATE()), 101)
EXEC sp_delete_backuphistory @DaysToKeepHistory
GO

-- When setting up SQL Server, it's handy to set the default backup location:
-- http://www.mssqltips.com/tip.asp?tip=1583
-- If we open the registry using REGEDIT or some other tool and if you navigate to this key:
-- HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft SQL Server\MSSQL.2\MSSQLServer or similar for
-- your instance of SQL Server we can see the following information The highlighted line below
-- shows the default directory for the backups for this one instance.

MultiChron 1.1 Released

2010-09-15T11:44:00.000-07:00

It's been a long road, but CaitSith and I released MultiChron 1.1 over the weekend on Adafruit. There are now 6 clocks in the MultiChron. CaitSith merged in his DeathChron variant, a count down clock, and I created a new TSChron, the TimesSquareChron. I've always been fascinated at how the time changes as it marches across the marquee at Times Square. Now you can see this same phenominom on your Multichron.

TimesSquareChron from Charley Jones on Vimeo.

We were also able to include a GPS module in the new code. If you attach a standard GPS module to the clock (instructions at http://crjones.com/adamods) the clock will auto update when it receives a fix.

I'd also like to thank CaitSith again for all his help on this. I wrote a lot of code, but none of it would work or even fit without all of CaitSith's very patient testing and tinkering. Thank you so much.

You can get the code at here:

The Project Source is here.

Ubuntu 10.04 Dell Vostro A90 : Reinstall Notes Sun Java, Arduino, Yahoo (Gyachi)

2010-09-03T23:09:00.000-07:00

Ok, ran across some more problems reloading the mini.

Sun Java Runtime:
It's suggested by several programs that you run Sun Java JRE, but you won't find it search for Ubuntu Software Center until you Edit : Software Sources : Other Software, Enable Cananocial partner, then install Sun-Java6-JRE. But Sometimes cut and past is easier:

sudo apt-get install sun-java6-jre

Arduino:
Looks like the install notes on Arduino.CC are getting a bit out of date. I think the thing that keeps confusing me is that gcc-avr appears to include both avr-gcc and avr-gcc++. I wish someone had told me that!

The pac-kages that you need to install to do command line compiles (at least) are:

sudo apt-get install gcc-avr avr-libc avrdude

Yahoo on Linux:
To install Yahoo support, it's gyachi. Until recently, these packages have been broken. Looks like baudm picked up where loell left off:

sudo add-apt-repository ppa:baudm
sudo apt-get update
sudo apt-get install gyachi

Ubuntu 10.04 Dell Vostro A90 : XFCE disappearing Menus and Panning the Screen (SetRes.sh)

2010-09-03T22:24:00.000-07:00

It was time to reload the mini the again, and the update went pretty well. I just wanted to document a few of the problems that I saw, so hopefully I won't have far to go to research them again in future...

Disappearing XFCE Menus:
First of all, XFCE still has an annoying way of losing the menus. Ya, ya, it's some keystroke and command, but the brain gets foggy...
ALT F2
xfce4-panel

SetRes.sh
Secondly, the A90 is only a 9" screen, which makes it adorable and portable, but very much a pain in the ass to work with on some programs. It's a full 1024 across, but only 600 pixels verticle. To scroll the extra 200 pixels that a normal screen would have... I created a script called SetRes.sh and dropped that in the ~ home folder. That script must be marked as executable and contains:
xrandr --output LVDS1 --panning 1024x800

Dataman merges branches of MonoChron

2010-08-07T14:34:00.000-07:00

Earlier this year I was the first end user to create an end-user hack for the MonoChron clock from Adafruit. The hack loads a clock that simulates a popular game from the 80's on this retro style clock intended for hacking:

Since then, there have been 2 more versions of the clock published by the manufacturer. Along with the original version, that's 4 different variants of the code. Quite sadly, if you wanted to change to any of the different variants it a rather lengthy process to connect cables and download code.

What's needed is some sort of crazy person with enough time and talent to merge all these different branches into one code base. Enter the Dataman.

I've been able to merge2 of the code branches right now, and just about 10k left over. I think I can cheat 2.5k of that back by shortening the bootloader down. And I think I can optimize the code further. So I think all 4 code branches are possible.

2 down 2 to go.

More to come.

Finally! FRF91 Rooted on Nexus One

2010-07-28T22:19:00.000-07:00

I finally rooted FRF91 on my Nexus One!

It's taken me about a month! Ever since they dropped the FRF91 update on top of us I lost root access to the phone. Wouldn't be so bad, but that's meant no backups, no advanced power control, none of the really good stuff. There was a patch of course, but it didn't seem to work. I must have tried about 10 times to reload the phone, step by step, all the way back to ERD79, ERE27, FRF50, and then then the attempt at FRF91.

Well, here's the formula. Flash back to ERD79. Let your phone do an OTA Upgrade to FRF91. It's that simple.

Then follow the instructions here:

Basically, you're going to load Amon-RA's Recovery Image. Then use that to load the damn FRF91 Root Update. BAH Hahmbug! What I was doing wrong was renaming the file. Damn thing didn't like being renamed! Only way to use it is with Amon-RA!

Spoiled MakerBot Brat gets het own Blog

2010-07-09T17:15:00.000-07:00

Well, I thought it was about time that I documented the progress on the Makerbot properly. Introducing DataBot's Blog at http://DataBotLV.BlogSpot.Com. Goes all the way back to the beginning, day by day...

Houston, We have Ignition! Makerbot is Online!

2010-07-07T02:51:00.000-07:00

My troublesome child, Makerbot 1591 finally relented into being a good girl and started printing out like mad. Very afraid to change plastic and mess up anything right now, so just going with the flow. Going to get all the upgrade parts printed, and who cares if they're white abs plastic! The Z-Axis lever is pretty cool, as are the slip on Z-Flags.

Some problems with the heat blanket interfering with the head. Will need to pull that back a bit tomorrow. In a stroke of brilliance, we coated one side of the thermal blanket with kapstan tape. You know you'll be taking if off again, why not make it easier and more stable? The tape should also add a layer of insulation to the blanket. Its a win win!

Here's the haul from tonight. All objects courtesy of Thingiverse.Com.

Continued Problems with Extruder

2010-07-04T00:18:00.000-07:00

Well, we learned the trick of heating up the extruder head to get the gunk out, and then soaking it in acetone to get the rest of it out. Problem is, Padawan tripped over the cat, and the hot head went tumbling. As luck would have it, it landed on the fragile and very hot tip which snapped off. So, we had no choice but to try removing the tip. This is the result.

That tip is never coming off, no how. Bummer. Anyhow, emergency kit was ordered 3 days ago. Should be here on Tuesday due to the holiday. Back in business then.

Makerbot Progress

2010-07-02T09:29:00.000-07:00

We’ve been having problems with pinchwheel, driverboards, and barrel assembly this week.

That’s down from the multitude of other problems we’ve overcome.

Pinchwheel:
The pinchwheel feeds the thread to the machine,
We’ve upgraded to a MK5 gear, that grabs the thread and doesn’t let go.
And we’ve become experts in taking it apart and adjusting tension.
I think we’ve nailed down this component now.
Unfortunately, it’s plexiglass, and we’ve broken it in 2 major places,
And fractured it in a few more.
Still operational, but for safety, we ordered 2 more.
Due here next week.

Driverboards:
The driverboards X, Y, and Z are fragile,
And we’ve learned how to work with them.
Rule 1: Don’t hot plug! It kills the driver chip.
Rule 2: Don’t overtighten drive belts.
Rule 3: Make sure all plugs are firmly plugged!
I blew up one driver board on Monday, got a new driver chip,
And had it up and running again by Thursday.
It’s surface mount, and not for the faint hearted,
But with a little ChipQuick, it’s a pretty easy fix.
Somehow, I think I'll blow one up again someday,
So for future reference, it's an Allegra A3982SLBTR-T,
Available at Digikey.

Extruder:
The Achilie’s heal of the whole system is the Extruder header,
Even with everything working 100%, this system will drag everything else to its knees.
This was our first build of the head, the floss enters through the left, the PTFE,
A sort of cool squishy barrier,
It then enters a copper barrel,
Is heated, pressurized, and finally exits the nozzel.

1) PTFE has become compressed and not allowed thread to pass.
We solved this by reaming out the PTFE with a 1/8” drill bit.

2) Barrel has become compacted with frozen thread not allowing more thread to pass.
We solved this by stripping the barrel bare, heating up to 220 in an oven for an hour to loosen the goo.
Then bathing it in acetone overnight to finish cleaning it out.

3) Thermistor has gone wild reporting invalid temperatures.
Still wondering out about this one. We plugged in one night and the temperature spiked to 255.
We pulled down, were heart sick that the thermistor had died, remeasured and remeasured again,
Finally got a reading. Resoldered. Been working fine since.
I put a couple on order at Digikey just in case.

As of this morning,
All problems are clear,
And looking forward to printing tonight.

But there again,
For the last 5 nights,
We've been clearing every problem,
And been ready to print.
That's when the new problems crop up.

But we've been ordering spare parts daily,
And now have a replacement part for almost every tiny breakable part.

Makerbot Build Day 1

2010-06-25T13:18:00.000-07:00

Well, this is truly the build of a lifetime. My padawan and I are building a Makerbot.Com for my ehem, (undeciperable) birthday. It's a standalone computer that prints 3d objects. It's about as close to a Star Trek replicator as we'll get in this century.

Completed the Y, X, Case, and End Stops builds tonight. Went fairly well except for a few mis starts. Lots of missing pictures in the build instructions. Sadly many of the pictures outdated or wrong. Wrong color parts and such. Oh well. All that much more challenging...

Some notes for future generations...

When installing the build platform... you have to flip over the final surface before inserting the magnets or the two platforms will repel each other, which is not good.

Be sure to read the instructions twice, or you'll be pulling stuff apart to put it back together again. DUH. So far, have had at least one duh error per build stage.

Stage 1: Magnets repelled on build platform. Simple, just unscrewed and flipped over the build platform.

Stage 2: Mounted the Y stage upside down on the X stage. Duh. Unbolted and switched it around. Watch that the text lines up. May not have made a big difference.

Stage 3: Missed the part about putting the top on. Duh. Had to back out two bolts. Locktite does not taste very good.

Here's a picture of the progress so far. XY stage is about where it's going to go. Case is assembled. Yah, I jumped the gun and attached all the electronics.

If buying this kit:

They do supply all the allen wrenches, ribbon cables, sand paper. Ok, damn straight up and nice of them.

Kit is designed for their specific power supply, with fits their power panel with custom printed serial number, and custom drilled specific bolt pattern. Buy theirs or you'll be cutting! You many not want to.

They don't specifically tell you that you need: zip ties, ethernet cables (7 if doing endstops, only 1 otherwise), a standard power supply cable. Annoying, but logical.

Suggested, Highly Suggested, Locktite everything. Get a little dixie cup, cut it down, make a little puddle of locktite, and dip each screw before placing it. We're building forever.

A tale of two makebots

2010-06-23T17:12:00.000-07:00

The toy of the millenia is on the way from NY. A Makerbot, in two boxes. One arriving today, the other, tomorrow. Then it's a very very long build to the finish.

If you've never seen it, it's a combination of really beautiful engineering, and somewhat ugly styling. Sorry guys, it's it could be cuter. Don't know. I might have to upgrade mine to clear plastic panels at some point.

The question you're probably asking right now is what the hell is a MakerBot and what is for. Well, imagine any object made from plastic, that measure under 4"x4"x6". This machine can make it.

Why is that exciting? Why wouldn't it be?

You can make anything you can dream of!

I've been chiding my padawan cobuilder for the slow order by drawing these cartoons. Its been a pretty agonizing week waiting on the packages.

We made the mistake of ordering many of the parts and supplies piecemeal to break up the cost. Padawan hadn't learned about 3 day ground from NY, otherwise it takes freakn ages to get stuff from New York.

Apparently the tomato truck the bot was on broke down or something on the way to Vegas. And the Bot had to hitch the rest of the way.

Box #2 was apparently on Rocket Rails and hit Vegas this morning....

LiveClock tops 500 downloads

2010-06-20T15:43:00.000-07:00

If you build it, they will come. And they have. 514 of'm in 5 days. Putting LiveClock in the 500-1000 Category. At least that was as of Saturday morning. It's come a long way since last Sunday night. I just posted the 5th update to the software. There are a total of 29 configuration options. Everything from hand start, end, width, color, and end style. I can't believe it's really come this far.

Just looking at the numbers right now, it's getting ready to cross 600 downloads. Of that, 340 people have actually kept the app on their droids.

Ya know, AT&T, you could have had this app on iPhone, for free, had you just been a bit nicer to me. ;-O

And the fortune cookie says:

2010-06-16T23:42:00.000-07:00

Had Chinese the other night, and no kidding, this was the actual fortune cookie I got. Well, I probably could have used this confirmation about 25 years ago, but better late than never...

LiveClock version 3.0

2010-06-16T08:48:00.000-07:00

3 Versions in 72 hours

Learning a new language (java) and a new platform (android) isn't a terribly easy thing to do. I've got to say that doing a serious project in java is a lot easier than I thought it would be. There's a lot of similarity to C++ and a lot of the same coding conventions work. Sadly, there's a lot of brain dead stupidity, like no unsigned integer type that I can find. I blew a day trying to stuff a color value into an integer, only to figure out, it was overflowing the integer. Oh well...

LiveClock has turned into a highly customizable Live Background. Leaps and bounds from where it started. People are taking notice. It's already topped 300 downloads on the marketplace since Sunday night. I even go a request last night from a Korean company to distribute the program.

I've been wearing a Movado watch for a number of years. That's where the idea for the clock originally came from. So only fitting that the Twelve Dot is the latest option, along with the option to set the size and color of the clock.

More to come.

Published my first Android App: LiveClock

2010-06-13T23:33:00.000-07:00

Published my first application for the Android. It's a Live Wallpaper which is a feature from the 2.1 Operating System. Basically it's a background for your phone that moves and changes behind your icons. I wanted something classy like a Movado watch, just simple, very plain. Just the 3 basic hands, hour, minute, and second.

Actually, this is the standard test application that I usually write in every graphic OS to make sure that I know what I'm doing.

Let me tell you what a pain it was to get this code up to the Droid Marketplace. Certificates, Encryption, Zipfiles, Makefiles, Icons, More Icons, Yet more Icons!

Anyhow, it's done, now free on the Android Marketplace. Probably another version coming in the future.

You can check a copy of the Android Marketplace posting here.

Panning Desktop in Ubuntu

2010-06-09T23:28:00.000-07:00

I've been trying to run the Eclipse as part of the Android SDK under Xubuntu 10.04 LTS on my Dell Mini A90. Problem is, the screen resolution is only 1024 x 600. That's great for most applications, but I've found a dirty little secret. Seems that some apps like Eclipse, trim their dialogs (in the middle!) in light of the shorter screen vertical. The specific problem was not being able to chose a platform (Android 2.2) when creating a new application.

So, what I needed was a way of behind able to expand the screen a bit, bigger than the physical size, and pan around a bit. It's called Panning, and the A90 has more than enough video memory to support it. Actually, it's pretty simple.

xrandr

Will report on current video modes and devices.
LVDS1 was supporting 1024x600.

xrandr --output LVDS1 --panning 1024x768

Tells LVDS1 to support a slightly larger vertical.

xrandr --output LVDS1 --pannning 1024x600

Resets display back to normal.

Loading Android 2.2 (Froyo) to Nexus One AT&T Version

2010-05-27T12:49:00.000-07:00

The problem with the leaked 2.2 Froyo update is that it's designed for the Netxtel version of the phone, not the AT&T Version. The update does a version check which fails.

The following rough procedure loads the Nextel Shipped image to the phone, updates it to the current version, then finally installs the Froyo update.

I'll update with better links soon. The main purpose is to remind me what to do.

This procedure ran on Xubuntu 10.04 LTS.

You need 7 things to make this work:
1) A Google Android AT&T Phone, will not work with Moto Droid.
2) Android SDK
3) Fastboot
4) Android USB Cable
5) ERD79 Base image
6) ERE23 Update
7) FRF50 (Froyo) Update

Install Android SDK
Download and copy fastboot to Android/tools
cd Android/tools

Make sure adb is running:
sudo ./adb kill-server
sudo ./adb start-server

Unlock and clear phone:
Boot phone by holding down scroll button & power on.
sudo ./fastboot oem unlock
Accept license, phone will clear user data.

Reboot phone by holding down scroll button & power on:
sudo ./fastboot erase userdata
sudo ./fastboot erase cache

Download and unzip ERD79 rom to Android/tools/79

Flash the E79 (Nextel Shipped) Rom:
sudo ./fastboot flash boot ./79/boot.img
sudo ./fastboot flash system ./79/system.img
sudo ./fastboot flash recovery ./79/recovery.img
sudo ./fastboot reboot
Let phone reboot a few times.
When finally booted,
Connect USB to computer

Download ERD79 to ERE27 Update
Copy to phone SD card as update.zip

Reboot phone to bootloader.
Click on Power Bootloader.
Volume Down
Click Power on Recovery
At Android with Exlamation Point,
click Power & Volume up at the same time.
May take a few tries.
Volume down to Apply Update.ZIP
Click Power.

Phone updates to ERE27 (Nextel).
Reboot phone to OS.
Connect to USB

Download FRF50 update.
Replace update.zip on SD card with FRF50 renamed as update.zip
Reboot to Recovery as before..
Apply Update.Zip as before.

Wala, an FRF40 (Froyo) AT&T phone.

Dataman finds bug in own programming....

2010-05-26T22:54:00.000-07:00

Well, I feel well smote right now. Contracted Bells Palsy a somewhat rare disorder that totally screws up your life by paralyzing half of your face. No ability to close your eye, it has to be taped shut. Not much ability to speak and eating is very difficult. Luckily, and I hope, this is a temporary setback. I have better use of the left eye now, thank God. I was going crazy of a week of mono-eye. Couldn't see jack with one eye! One eye'd Chuck? Ok, God, point taken, slow down. Gotcha loud and clear.

I'm OK. Speech is still a bit distorted by a lagging upper lip, but otherwise, the left side droopiness is going away. Slowly. Way too slowly.

Datamans solves Game of Life

2010-05-05T15:56:00.000-07:00

I was having some problems with the Game of Life kit from Adafruit.Com. It's a great kit. You're able to build modules of 16 leds, then plug the boards into each other to create larger arrays. Problem is, in some configurations, the boards occassionally lock into oscillating patterns. Some as simple as frequencies of 2, some as complex as fequencies of 16. What I really wanted was the panels to remain random and self reset should they see a repeating pattern. I've worked on the code for a while and posted a solution to this problem on GitHub.

Installing Lucent Linux on Dell A90

2010-04-30T14:26:00.000-07:00

I'm really loving xunbuntu 10.04 on my Dell A90. Thank you so much XUbuntu Team and the entire Ubuntu community. You have made a believer out of me. You can get your free copy of XUbuntu (comparable to Windows Home Edition / stripped down) at http://xubuntu.com/. You can install more packages through the feature rich interface to bring it up to full ubuntu (core), or simply install one of the other free versions: kubuntu (kde), edubuntu (educationa), server, mythububuntu (home theater), or ubuntu studio (multimedia). More at Ubuntu.Com Derivatives.

There have been 3 pretty consistent problems in getting a Dell A90 to work with Ubuntu. I found this great posting at http://www.wootens.net/?p=975 that reminds me of how to fix them every time. While not completely accurate, it’s a great start:

NO WIRELESS:
… you have to get wireless network to work. Using the wired ethernet, do the following:

sudo apt-get update
sudo apt-get --reinstall install bcmwl-kernel-source

(I’ve found that all I had to do was enable the proprietary STA driver -crj)

NO SOUND:
Next up the sound will not work (LAME!) Do the following:

sudo nano /etc/modprobe.d/alsa-base.conf

Add the following at the end of the file:

options snd-hda-intel model=dell

(I’ve found I’ve always had to do this to get sound to work, Additionally, make sure that speaker volume is cranked up in sound control panel! -crj)

SLEEP HANG:
And finally the machine will hang if there is an SD-card mounted when the machine sleeps. This requires the following hack:

sudo nano /usr/sbin/pm-suspend

add the following to the beginning of the file:

# HACK - WLW umount sd card!
umount /media/

(I’ve found I’ve always had to name the SD card, ie: umount /media/DATA where DATA is the label of the SD card - crj)

2010-04-15T11:44:00.000-07:00

Newest toy is a Teensy: about the size of a penny and a half. Arduino compatible. Built in mini-USB, Led, Crystal, and reset button. Under $20. In fact, it's so small, you could probably stack 2 of'm in about the size of the male USB A side connector. Looks like the thing doubles the I/O of the 328... No FTDI chip, stack is build directly onto the chip. Can't wait to start playing with this thing.

Epiphany, a framless laser cutter, cnc, extruder...

2010-04-05T01:43:00.000-07:00

I've got a crazy idea to make a home laster cutter/etcher.

From what I see there are 3 axis to CNC,
X, Y, Z.
Every axis you can cut out reduces the cost.

A laser cutter controls Z through the power of the laser.
All that's left is X and y.
X is pretty simply the distance left and right on a beam,
Y is typically trickier.
It's gotta be along a second rail.

Lets consider a device for a moment that's a laser cutter at its
simplest. A rail with 4 wheels for feet. You would place this cutter
on a pad on a table, it would roll itself on the Y plane using very very
simple stepper motors. 4 wheels, no choice but to travel in a straight line.

That leaves X, the rail. Typically this is long screw hooked up to a
stepper. Or something similar to printer.

Shroud the thing in dark plastic to protect our eyes from the laser.

So, it's a little robot box that moves on a table. Simplest form of
laser cutting possible.

Electronics package would be, control for x, y, power for laser, and usb
serial i/o.

I don't think this can get much simpler.

The beauty of the thing is the size. we build it to read a piece of
paper, a safety blotter on the table. When it gets to either end, it
stops.

Rather than having the expense of a frame for the laser, a box for it to
sit in, we use the existing environment of the desktop and make it safe. One with the environment. very zen.

Ok, lets take this one step further.

If this same table top bot can cut,
Why can't it extrude as well?
Could this be a seriously simple tabletop extruder?

Which would mean the Z axis would need to travel
But hell, same kinda jig.

Will have to draw up some sketches.

Uhm, chucky, stop being stupid.
Simplify yet again.
Why are you thinking 2 dimensional.

How about table top bot that carries around a laser, extruder, whatever.
And simply takes commands to move X, Y steppers.
No screws, just a simple bot, that moves over the surface doing its work.
An umbilical feeds it instructions, power, material.

Er, isn't that called a mouse at that point?
Hmmm, need to work out wheels that can move both x and y.
Steppers...
Steppers and wheels I've got.

Oh, the madness...

SMT Boards going strong

2010-04-03T12:18:00.000-07:00

Have been working hard on the SMTBoards.Com website and preparing inventory. Not easy. I started a blog on SMTBoards, and have been blogging about projects there. http://smtboards.com/

Tiny13 Programmer done!

2010-03-31T12:33:00.000-07:00

I promised to post instructions and kits on my Tiny13 programmer, and I got it all done last night. This is the same basic design I used to program all the SMTCylon CPUs. This version is a bit kinder with larger traces to drop the chip on. Where I used to have failures to program about 1 time out 3 with the previous board, I'm 0 failures so far on this one.

I reworked the SMTBoards.Com web site for a more professional look, gave it a shopping cart, and lots of goodies in the shop. Quite sadly, the shipping calculator is off, and charging a bit too much for shipping. I only charge cost for shipping, and will refund any difference.

I also decided that SMTBoards needs its own blog, and started one over there. It can be reached easily from the Blog menu of SmtBoards.Com.

SMTCylon - Better Packaging

2010-03-29T08:58:00.000-07:00

Whew!

The new sealer came on Friday and what a difference in packaging.
Here's what I started with:

And here's the new package.
Everything is separated.

I think that's a lot better.
It's 4 presses of the sealer,But doesn't really take that much time.

Constructing the kits directly into the bag also resolves a bunch of problems.
* Lets me visually inspect the package before it ships.
* Isolates the battery from the components.

I got boards back on Saturday, electrical tested, all looks good. A high number of failures due to bad exposure. Still learning the new exposure light. At least the boards are developing consistently.

Programming the CPUs didn't take that much time, and adds a lot to the kit. According to Dale, "How did you program these parts? SMT sockets are crazy expensive." Introduction my low tech TinyISP programmer. Drop the chip on the pad and clamp with tweezers. 50 more SMTCylons just joined the universe. I'll be posting up instructions for my TinyISP programmer shortly.

Kits in the Mail

2010-03-24T14:13:00.000-07:00

SMTCylon Version 1.1
Actual size 1 1/4" x 1"

Huge grin!

The first SMTCylon kits to Beta Testers are in the mail! Sent off five kits to my industry piers for comment, including the orignal designer, Dale Wheat. Can't wait to hear the feedback.

First commerical order already came in over the website. http://smtboards.com/smtcylon I'm now a published kit maker!

SMT Boards Progress

2010-03-19T09:41:00.000-07:00

Received my shipment of parts for SMT Boards last night. Decided to focus on SMTCylon as the first production project.

Christmas II
(for circuit geeks)

So, first question, does the Tiny13 cpu fit the circuit?
Yes, just barely...

Does everything else fit?
Oh yea...
For scale this whole circuit is 1.25" x 1" in size.

Ok, Geek, can you make tiny LEDs light up?
Check...
For scale again, that's a .1" connector in the picture.

Wohoo, and we already know how to burn printed circuits!
(Apparently we don't, 3 more failures...)
(Having problems getting the image to develop.)

But then I did get the image to develop,
New developer, New batch, Hot developer:

Well, not so lucky tonight.
The image quality was poor.
Inkjet did not print dark enough,
So there were gaps in the traces.
I saw the poor quality in the photo,
Thought it might fly, I was wrong.

All is not lost.
There's a toner transfer method that I can try with failed photo transfer boards.
The boards are still good and usable.

Need to improve my image quality...

Getting bold, going commercial

2010-03-17T14:59:00.000-07:00

Going to give Open Source Marketing a try. Open Source is a brave new world in the nerd world; You build something then you release it to the world for fee. Hopefully in doing so someone will pick up the thing you built and make it better.

Second piece of data: The majority of electronics kits created are through-hole by design. The components go through the holes in the board. On the other hand, Surface Mount (SMT) is cooler, smaller, requires a bit more skill, and there's not a lot of kits out there.

What if we combine the two? Take open source designs, usually through hole by design, and turn them into surface mount kits.

http://www.smtboards.com/ is born.

This is the through hole version of Dale Wheat's Tiny Cylon, an open source project:

And this is my SMT Version,
Still tiny, but battery onboard!

Parts are inbound, due Thursday!

Will cut a few sample boards tonight.

Learning to burn PCB boards

2010-03-11T12:46:00.000-08:00

Well, I thought I'd take the plunge and learn how to burn my own pcb boards at home. Every time I've tried to do this in the past, I've ended in failure because the costs have been astronomical. I ask the experts and get really turned off.

Adafruit blogged about a great article on Make about etching at home. The orginal article is here: http://blog.makezine.com/archive/2010/02/circuit_skills_circuit_board_etchin.html

Basically what it says is that this process can be done with very cheap supplies for under about $50. In reality there's a few more basic things to buy, the most expensive being transparency supplies for your inket or laser printer. It's a great article though and really spurred me on to try it.

This was the very first board I with my padawan. Note that the master ignored the comment from the padawan to leave it in the acid longer, and I ended up with a lovely circuit on a thin layer of copper, so the whole circuits conducts everywhere. Note to read, leave the board in the etchant till you can see through the board. The pcb is nearly transparent. Still pretty, but now it's an artifract of a failed attempt.

Same design, but now left in the etchant much longer. Note that we are seeing red ink from the underside of the board. Reads NP.

Cutting apart was easy with a set of 4" tin shears. Cut right through the board to separate. Next up was drilling with my Dremel Drill Press. This is simple a Dremel tool pluged into a stand, and works really well. With sizes so small though, a good set of magnifiers is needed.

Once it's drilled out time to solder... Good pick of me soldering... Of course, this was on the horrible first board that I didn't realize wasn't fully etched...

So, I tried to get fancy after that, add a ground plane to the board so that it doesn't take as much time to burn, use as much etchant, and reduce signal noise.

In theory, that workds pretty well. In practice, I didn't give myself enough clearance between the drill holes and ground. Notice that the drill holes, although pretty good, still done by hand, stray into the ground plane area. If I were to solder this, solder would definately bridge onto the ground plane.

Pretty looking board though. Just need to widen that gap...

The end result? Well, I did manage to drill out one board and produce a Multi ICSP header for programming 8 (Tiny13), 20 (Tiny2313), and 28 pin (AtMega328p) Atmel CPUs. That's good, because I'm getting tired of building up protoboard circuits for doing this. I used standard chip sockets and they're not really suitable for tons of insertions and removals, but good enough for a cheap programmer.

The multiICSP, not exatctly error proofed the first time. Had to hack the prototype a bit to make it work. Poor choice of ICSP connector (2mm vs .1") and hacked in an ocsiclator option.

The second project on that board was the Digg Clock. That's Lady Ada's Digg Board with a few additions to turn it into an accurate clock. Part II of in house board making is that you're prone to making mistakes. I used the wrong header, had to cut traces that were too close, had to reroute over cut traces, uhm, put the battery holder in backwards, didn't leave enough room to plug in a usb cable...

And quite sadly, this circuit does not work in any way intented. Too many mistakes. Luckily most of the expensive parts are recoverable, least I socketed the chips!

Dataman credited with first hack to Monochron

2010-03-05T10:14:00.000-08:00

Another post about this great company in New York, Adafruit. Two serious geeks having a blast creating and shipping hackable kits all over the world. http://adafruit.com/

Well, they prereleased code a week ago and I immediately started an "Alien Intruder" hack for the clock. I was able to debug the code in less than 24 hours after building the clock, and was credited with the first official hack to the clock. http://www.adafruit.com/blog/2010/03/04/invaderchron-a-monochron-hack/

Code is free to the world at http://github.com/dataman/IntruderChron

New clock

2010-03-02T23:40:00.000-08:00

I probably never posted that I built my first clock from Adafruit.Com, but I just built my second. Adafruit is a great little company selling open source kits. Open source means they publish all their plans right there out in the open. If you really wanted to, you could build any of their kits without any parts from them, even load their software onto it for free. But there again, they do sell all these cool kits.

The clock on the top, and the first clock I built from the them, is a Russion Ice Tube clock. Fantastic clock using surplus vacummn tubes directly from mother Russia. A little hard to see, but that's a clear acrylic case its in. Part of the brilliance here is a pump up circuit that pumps 5v up to 70v to power the tube. It dims itself at night, and is bright during the day.

The clock on the bottom is the newest one, and yes, it is Pong! The clock plays pong all day long. When someone misses, the time increments. So many other possibilities!!!

Best part of open source, it's all hackable. I've already updated the ice tube clock with new firmware. And I've started work on an upgrade for the Pong! clock.

Christmas II

2010-02-11T14:30:00.000-08:00

It's been Christmas II these past 3 weeks. Games that should have a been released for the Christmas season has short delays, and made the last three weeks Christmas II. So my play list is currently looking like:

Mass Effect 2	Just found the Collector Ship
Star Trek Online	Grins, my Klingon just took out a Star Fleet Constitution Class!
Bio Shock 2	Rescued my first Little Sister, Just took out a Big Sister!

Combatt 2.0	20 patches to test and playtesting!
Arduino	Panel8x8 projects, Jeelinks, and Clocks

And hey!
It's a 3 day weekend coming up!!!
Wohooo!!!!

Beam me up Scotty! Now Please!!!

2010-02-10T14:25:00.000-08:00

I beamed into the 23rd century! And I'm home! Been playing Star Trek Online for just about a month now, and finally made it to the level 6. What's so special about level 6? Well for one, you get to play a Klingon now, and take part in PvP (Player vs Player). Klingon vs Klingon, or massive fleet actions of Federation vs Aliens. Playing StarFleet is great and everything, but there's a certain beauty to being Klingon!

It really is a Trekie's dream. Fly around on the ship of your choice, beam down onto planets, kick alien ass, pet some fuzzy tribbles, buy some cool weapons. Just haven't found any of the green alien babes yet. Still looking...

Dataman meets the 4th Dimension

2010-02-04T11:13:00.000-08:00

Augmented reality is finally here. Well, not that it hasn't been here before, but it's finally here for real!

The first time I remember seeing Augmented Reality was in the movie FutureWorld. This was the sequel to WestWorld, remember Yule Brener as the Cowboy in Black? Well, they had this chess set where the characters came to life. But that was just a movie.

The chess game played in Star Wars (above) was in 3d. Just took us a while to catch up to the movies.

I bought the PS3 specifically to play Eye of Judgement (right), and it's pretty cool technology. The computer recognizes cards thrown down on the play mat and produces 3d images on the screen, ala Star Wars Chess Set. And that game came out in 2007!

Well, last night I was playing around with my iTags again. These are included with all of the Avatar action figures. I tried the software last month from AvatariTag.Com, and it crashed my browser. Looks like they've gotten it fixed.

Show the tag to your webcam, and the computer animates a 3d model with sound. You can rotate the model 360 degrees and even tap on an info icon to bring up information. Notice there's every a little guy inside the cockpit.

Well, that's November 2009 technology.

Last night I was cruising the net and came accross Harry Potter Orlando Preview. http://www.universalorlando.com/harrypotter/ You can print a map, and show it to your webcam to get a 3d augmented view. Worked fairly well as well. Great software install.

Only real problem is the integrated webcam on the laptop is a pain for this. To get a good camera angle on the object you have to pretty much block your view of the screen. I'd suggest using an external cam.

Ground Control to Major Tom!

2010-01-21T15:12:00.000-08:00

Mars Odyssey giving up the search for Mars Phoenix for now. The belief is that the little guy probably got crushed in tons of ice. A fitting end for a poor little probe sent to find water. Odyssey to try again in February and March.

Mars Rover Spirit is still stuck in the sand on Mars. Sadly, if they can't get the solar panels orientated correctly before Martian Winter sets in, it's doubtful we'll hear back from Spirit in the Spring either. Opportunity is till chugging along strong though.

All in all, these little probes did a great job. Designed for 3-6 months, the Rovers have been on Mars for more than 6 years. In the case of Phoenix, it was risky landing a probe in the polar cap, and it was never expected to make it through Winter.

I found this cute picture looking for a thumbnail for this post. If only it were true...

Auto Reconnect in XNA

2010-01-08T10:09:00.000-08:00

Had a problem the other night when were towards the end of a network game in Commbatt. Suddently the "Comm Error" message appeared and the game was over. Damn, can you believe we can't recover from a simple comm error?

Wait a minute... I already know the game handles of the players. None of the data has really changed. How about, I drop the network session, create a new one, rejoin the players, and resynch the maps?

Believe it or not, it works. I haven't worked out the resynch code yet, but that shouldn't be hard. But I did work out the reconnect code. Worked like a champ. I really need to do some debugging on all paths, but I think it's viable. Maybe this is something the real games do in the background?

Just some thoughts:

Maybe I don't really need to drop and recreate the session on the host? Maybe it's just the player that needs to try and reconnect. I think I may have problems with AllowJoinGameInSession, or whatever it's called. Will have to look at that.

How do I keep other players from joining these sessions looking for a rejoin. Seems that they appear in the standard available games list. If a new player did manage to sneak into an existing session, there wouldn't be enough data to host the new session. (Each system know the object and coordinates, but not their inventory.)

Testing late last night I think I saw a host migration. My test account had dropped and reconnected to the server. I tried again, and this time it was trying to connect to itself? Just needs looked into.

Wierd behavior in Oracle Triggers

2010-01-08T09:43:00.000-08:00

A fellow DBA at work was having some problems with a BeforeUpdate Trigger in Oracle. His trigger was firing, executing the one function properly, then not performing the actual update as originally intended. Digging through the web, he was able to find an obscure reference that Triggers must assign at least one value, or the intended action, the update, will not work. The solution was to create a local variable and assign it to null. Seems like this is something that should be done as a standard in each Trigger just in case.

create or replace trigger trg_beforeupdate_table
before update on table
declare
dummy number;
begin
dummy:= 1;
end;

Creator loses match to Beta Tester

2010-01-07T11:48:00.000-08:00

Started testing RC1 of commbatt and noted just a few minor problems. Mainly cosmetic. A change to menu system now allows unlimited columns. Optional vertical lines for the columns can also be drawn. I used this new functionality to create built in debuggers for the messaging system and the player database. I may publish asccess to these debug windows in the single player game.

One bug, which is a bit vexing, is that we encountered a communications failure toward the end of the game. Not a game fault, but a bonnified communications failure where packets were not received in a timely fashion. With no way to correct the error, the game is over, and that's it. A lot of work down the tubes. I think I've come up with a way to renew the connection, which will make the game more resiliant to these types of errors. I'll have to resynch the databases, but I think recovering from a communications error will be worth it.

The client is stable enough that we got a few good games in. My tester is really coming along and showing some teeth. In fact, I lost a game this round. It's good to be humble.

Well, back to coding I guess. Going to recode the connection system and add some period sound effects to the sequence.

CommBatt 2.0 Playtesting

2010-01-04T13:02:00.000-08:00

CommBatt 2.0 spent the last week playtesting on Creators.XNA.Com. Sadly, not a single person reviewed the game. My fault really I guess, I don't spend much time there reviewing games as well. But still, not a single comment. So, I've vowed to comment on any playtesting games with low feedback. There were a few zero comments today and I took care of that. I guess the playtest community is as good as you make it.

Been testing with friends all week, and that playtesting really brought the game forward by miles. It's not quite the same game it was a week ago. It's matured.

A while ago I rescued a client from a vicious virus, Perfect Defender I think. Was a horrible Scareware virus that scared the person into buying an antivirus program, which itself was the virus. Anyhow, this guy turned out to be a lead graphics designer, and hence, the favor of box art as pay back.

Anyhow, I thought I'd share their initial design art. It's really good for the right game, but not quite for CommBatt 2.0.

Actually, so good that I may just have to write this version of CommBatt 2.0 as well...