Well the good news is that the issue causing SQL Server 2005 to restart itself is fixed in SP3. The bad news is that I do not have a test server to test the upgrade on. While working to get a test server setup, I decided to create some SQL Server Alerts. At least then I will know just a few moments after the event happens. Currently I would find out about the error if users reported a problem or during my daily review of the SQL Server Logs.

Here is how I setup SQL Server 2005 alerts.

1. Setup Database Mail. Just follow the wizard it will lead you to safety. When finished you can right click on Database Mail and Choose Send Test Email to make sure you have configured it properly.

2. Right Click on Database Mail and click Configure. Choose the Manage Profile Security option and click next. Make sure your public profile is set as the Default Profile.

3. Setup some operators. Give the operator a name and an email address at minimum.

4. Create your new alert. Right click alerts under SQL Server Agent and pick new alert.

• Give your alert a name.
• Under the Alerts will be raised based on section choose the error number.\
• In my instance the following error message would signal the service restart: Error: 3449, Severity: 21, State: 1 So for error number I typed: 3449
• On the response tab check the Notify Operators and choose who gets notified.
• On the Options tab check the include alert error text in Email option. Click OK. You have now created the alert.

5. Right click on the SQL Server Agent and choose Properties. Click on the Alert System option and enable the Mail Profile. Choose Database Mail for your mail system and then choose the profile you setup in the step above. Click OK. You will then need to restart your SQL Server Agent Service.

There are tons of performance conditions and error messages you can monitor using the method above. You can also set alerts up to fire on you SQL Server Jobs. Handy if you have a job that has started failing and you want to keep a close eye on it. However if you have 100s of jobs to monitor you probably want to setup some kind of Centralized Management Server to monitor that. You do not want to go add alerts to every single one of your jobs.

Once you have the Database Mail, Operator, and Alert system setup you can explore the space and use it for all sorts of monitoring scenarios. Run wild.

Update: You can use the RAISERROR system function to test your alerts.

The more I learn about the ethos of social media, the more baffled I am when I read examples like this one that Emily recently blogged about.  I mean this is Johnson & Johnson making this mistake!?  I guess in fairness it’s important to mention that ad company Taxi  was the one responsible for the Motrin account, but still? 

What’s baffling is the tone of the ad not the fact that Twitter became a soapbox for offended moms.  I understand where the company was going with this ad.  I have “worn” two babies in a Bjorn, and yes, I often felt like I needed Motrin, and lots of it, after “wearing them” for long periods of time.  However, some think this is no laughing matter, check out Babywearing International, Inc.  This was so avoidable in my opinion.  Johnson & Johnson had established relationships with plenty of moms.  Not four months before this campaign was launched, they were courting mommy bloggers at a Johnson & Johnson Baby Camp event. Johnson & Johnson is known for being a family friendly employer and in 2008 was an eight-time honoree on Fortune’s “100 Best Companies to Work For” in America. They, meaning Taxi, should have utilized these relationships and tested the ad with moms first.   

Also baffling is that Taxi wasn’t monitoring the mayhem.  According to blogger Joyce Schwartz, when she reached Taxi’s Director of Corporate Communications to get a comment on the over 2000 blog posts about the Motrin ad, she had to explain what Twitter was to the ad rep. In 2008, there was an estimated 4-5 million people on Twitter, and about 70% joined that year.  I still don’t find it reasonable that this person at the ad company didn’t know about Twitter and therefore wasn’t monitoring it on behalf of clients.  According to Taxi’s current client list, they are still managing the Motrin account. 

Go figure?

What is Zabbix ?

http://www.zabbix.com/

Installation on Mandriva 2009.1 Spring

1. Adding repository arsip.or.id

[root@srempuan telo]#urpmi.addmedia –distrib –mirrorlist http://www.arsip.or.id/mandrivalinux/official/2009.1/i586/

2. Install Zabbix Server

[root@srempuan telo]# urpmi zabbix

To satisfy dependencies, the following packages are going to be installed:

Package Version Release Arch

(medium “main_rilis_arsip”)

fping 2.4b2 12mdv2009.1 i586

(medium “contrib_rilis_arsip”)

libiksemel3 1.3 4mdv2009.0 i586

zabbix 1.6.2 3mdv2009.1 i586

2.7MB of additional disk space will be used.

418KB of packages will be retrieved.

Proceed with the installation of the 3 packages? (Y/n)

http://telkom.arsip.or.id/mandrivalinux/official/2009.1/i586/media/main/release/fping-2.4b2-12mdv2009.1.i586.rpm

http://telkom.arsip.or.id/mandrivalinux/official/2009.1/i586/media/contrib/release/libiksemel3-1.3-4mdv2009.0.i586.rpm

http://telkom.arsip.or.id/mandrivalinux/official/2009.1/i586/media/contrib/release/zabbix-1.6.2-3mdv2009.1.i586.rpm

installing libiksemel3-1.3-4mdv2009.0.i586.rpm zabbix-1.6.2-3mdv2009.1.i586.rpm fping-2.4b2-12mdv2009.1.i586.rpm from /var/cache/urpmi/rpms

Preparing… ##############################################################################################################################################

1/3: fping ##############################################################################################################################################

2/3: libiksemel3 ##############################################################################################################################################

3/3: zabbix ##############################################################################################################################################

[root@srempuan telo]#

more…

Download source code

Today we are going to create a windows service monitoring application in C#.Net.

A windows service is an application which runs in background without a GUI. They are used to monitor various activities such as database operations and message queues. For an example a windows service may periodically monitor a message queue and if new entries are added to the queue, the service will treat them. By using Visual Studio, creation and deployment of windows services is extremely easy. After deployment of the service, the services window comes under Administrative tools can be used to start, stop, pause and resume the service.
Today we are not going to discuss about creation of windows services. We focus on creating an application for monitoring any windows service. This includes both local and remote services! This application will reside in the system tray, connects to any service in the network and displays the current status of the service. Let’s create the application.

ServiceController class found in System.ServiceProcess namespace can be used to monitor services. We can create an instance of this class and get the current status from the Status property. The code for this is as follows;

System.ServiceProcess.ServiceController sc = new ServiceController(service, machine);
ServiceControllerStatus newStatus = sc.Status;

Here, service variable contains a service name and the machine variable contains a machine name.

We have to create a method which polls the status periodically without stopping. For that, a while (true)construct is the way to go. Since we should be able to connect to multiple services, we keep the settings in a text file to change/read them at will. Also this service monitoring section has to be executed in a separate thread other than the main UI thread because a context menu would be accompanied with the application and it will work as part of the main thread. The method having this code would be as follows;

When you go through the above piece of code, you may be confused by the lock keyword and the notifyIcon1. The lock keyword is used in order to access shared resources in a thread safe manner. Here the settings.txt file may be accessed by this monitorStatusChange() method and the UI thread simultaneously. For an instance, the context menu will have a menu item called “Manage settings” and by clicking on that, a form will be displayed to edit the settings. At the time of saving the modifications by this form, if the monitoring thread tries to read settings from the same file, an IOException will be raised as follows,

“The process cannot access the file because it is being used by another process”

Notify icon is the icon which is visible in the system tray. As per the above method, we are changing the icon depending on the current status of the service.

I hope you are clear with the monitoring part. But you may not have written an application which resides in the system tray and below steps are given for creating such an app.

1. Add a windows forms application project for the solution.
2. To the default page, drag a Notify control and a contextmenustrip control.
3. Assign the contextmenustrip control to the Notify icon’s ContextMenuStrip property.
4. Make Notify control’s Visible property to true.
5. Select an icon for the Notify control’s icon property.
6. Change form’s Opacity to 0%.
7. Select form’s ShowInTaskBar property as false.
8. Select form’s FormBorderStyle property as fixedtoolwindow.

When running the application, you would see it as in the following image. Look at the settings window and the context menu above the System Tray icon (The particular service is in the stopped state here).

I hope you have got a clear idea about this application. The source code for the application can be downloaded from the link at the top. Remember I have used third party GUI controls for this application called DotNetBar controls. This is my first ever blog entry and you are warmly welcome for putting comments here.

I will be writing another blog post very soon for enhancing this application in order to control the service both locally and remotely.

Cheers!!!!

@hrouda shared a presentation by Jeremiah Owyang on Twitter, about the different stages of listening.

I spend alot of time doing exactly this and thought it worth sharing as it is short, simple, and gets to the point!