Last week recap and links:
Image courtesy of kanate / FreeDigitalPhotos.net
What are your best reads this week? Leave them in the comments below.
Consider the following code in HelloWorld.API.dll.
namespace HelloWorld.API {
public class Praise {
public static string Someone(string nameOfPersonToPraise) {
return string.Format("{0} you're awesome!",
nameOfPersonToPraise);
}
}
}
Make the API available with the Add-Type cmdlet. This makes all the types in a .NET assembly available in the current session.
Add-Type -Path PATH_TO\HelloWorld.API.dll
Now the static method can be called by using the :: (two colons) syntax.
[HelloWorld.API.Praise]::Someone("eric")
The output is
eric you’re awesome!
The output can be further processed by using a variable to hold it.
$praised = [HelloWorld.API.Praise]::Someone("eric")
$praised.ToUpper();
The output is
ERIC YOU’RE AWESOME!
When developing remember that the assembly is locked after the Add-Type until the session ends.
We use this to setup our automated (test) environment with an API. No need for custom console applications, just plain powershell.
This works for all .NET assemblies. Use it wisely.
The latest software update of Synology added the option to backup to Azure Storage. Here is my experience of setting it up, the first backup and the costs.
I created a storage account and enabled minimal logging of the blob storage. The default geo redundant replication option is good enough for my backup. After copying the secondary access key to the clipboard I logged in on my Synology to configure the backup.
In DSM, the software of Synology, I created a backup destination to Azure Storage. Then I created a data backup task to backup all folders to Azure Storage. You can combine these steps by creating a destination while creating a task, but I did this in two steps. The wizards in DSM makes this super easy.
Since I enabled logging of the blob storage the Azure portal showed a graph. Below a screenshot during the first backup. The green line is not the transfer speed but the number of requests. The smaller the files the more requests, because more files are uploaded. (Bit rate Wiki)
The total of 1TB needed to be uploaded over my 90 / 9 Mbit/s internet. It took about 12 days to complete. In these days the Synology remained responsive. I could stream movies, download stuff and print with the virtual AirPrint service.
Once the first backup is completed only the changes need to be uploaded. I’ve scheduled the backup task every week.
Azure Storage costs are the disk space used (€0.0358 per GB per month) and the traffic out. The traffic in is considered free of charge. In my case the 1TB backup costs around €36 each month. The costs will increase when I add new files.
The same storage in Amazon Glacier would be €11, but that isn’t part of my MSDN subscription 
Data backup on synology website
About Storage Analytics Metrics on msdn
Last week recap and links:
Image courtesy of kanate / FreeDigitalPhotos.net
What are your best reads this week? Leave them in the comments below.
In our project we group multiple machines and request information from each one. Sometimes this takes very long.
All machines are contacted in sequence. When one machine is offline the request will timeout in 10 seconds (default WCF behavior) This way the user must wait very long when some machines are offline. The graph below shows the messages captured with fiddler when one machine is offline. Look for the exclamation mark.
We could call the services asynchronous but that would mean handling events and messing up the code.
Using tasks we now call the services in parallel. In the code sample below are the old and new sources.
// old code foreach(var machine in machines) CallTheService(machine); // new code var tasks = machines.Select(machine => Task.Factory.StartNew(() => CallTheService(machine)) ).ToArray(); Task.WaitAll(tasks);
This way we don’t reduce the wait for an offline machine, but make better use of the idle time by doing other requests. The graph below shows the messages captured with fiddler of the same group of machine as the graph above. The total process took 16 seconds in stead of 35.
When a user has to wait it is always to long. Make sure you do the best you can to reduce the wait.
.NET Development by Eric 