• It’s easy to forget to check whether there are new versions.
• Solutions composed of packages output by different repositories.

Some of our applications have code distributed across different repositories and packages. Since these are different solutions its difficult to use existing tools to detect when these packages are getting out of sync. NugetXray has commands to do just that.

It’s available at nuget.org.

nuget install NugetXray

There are currently two commands: one to find out of date packages and another to find package version mismatches.

Find packages with newer versions

To generate a diff report:

.\NugetXray.exe diff -d C:\git\source\
Scanning https://www.nuget.org/api/v2 for packages.configs.
WindowsAzure.Storage.2.1.0.4                                | -5.0.0      | 3   configs
Newtonsoft.Json.6.0.8                                             | -3.0.0      | 4   configs
WindowsAzure.Storage.4.3.0                                   | -3.0.0      | 1   configs

This shows us how many major, minor or patch versions we are behind by. You can generate a report across multiple repositories by setting the directory parameter to the root of both repositories.

You can also output either json or html if you want to build further tooling around it. The html report can included as a third party teamcity report.

.\NugetXray.exe diff -d C:\git\source\ -f html -o diff.html

Find packages which need consolidating

To find whether you have multiple versions of the same package, you can run the duplicate command.

.\NugetXray.exe duplicate -d C:\git\source\
Scanning C:\git\ConveyorBelt\ for packages.configs.
WindowsAzure.Storage                                                   | 2 versions
Errors:   1

As before you can include multiple solutions under a single root directory to find duplicates across solutions. It also supports both Json and Html output.

Fail the build

If you want to integrate it with your build process, and fail when there are packages which need updating or consolidating, you can use the exit code. If calling from PowerShell you can get by calling $LASTEXITCODE. A negative exit code is returned to indicate there are results.

.\NugetXray.exe diff -d C:\git\source\
$LASTEXITCODE
-1

Verbose flag

If you want more details output around the failures and which packages.config they refer to you can specify the verbose flag.

If you have any feedback or feature requests you can create an issue or reach me @naeemkhedarun.

As a stop-gap to Octopus hopefully adding this feature I’ve created OctoSearch. This allows you to download and cache the variable sets with their variable collections locally as Text, Json or Html.

OctoSearch itself is a package available from nuget. It has been compiled down to a native executable so a dotnet core installation is not needed.

> nuget install OctoSearch
> cd .\tools\OctoSearch*\tools
> .\OctoSearch.exe

The first step is to login with the octopus server so we can create, download and cache an API token. This will be used for subsequent calls to octopus.

> .\OctoSearch.exe login -l https://octopus/ -u username
Please enter your password...
*********
Successfully logged in.

Now that we’re authenticated we can download and cache the variable sets and their variable collections. This cache will be used for our searches to reduce the load on the Octopus server. Variables marked as sensitive won’t have their values downloaded or cached; their variable names will be searchable but not their values.

> .\OctoSearch.exe cache
Saved LibraryVariableSet1.
Saved LibraryVariableSet2.
...

With the variables cached locally you can run fast searches and regenerate them into either Json or Html documents. To run a basic command line search you can use the search verb. It takes a regex so you can pass in basic text or more advanced text searches when you need to.

> .\OctoSearch.exe search --regex connectionstring
Database.ConnectionString            ConnectionStringOne
ServiceBus.ConnectionString          ConnectionStringTwo

To output the search results into a text file you can do:

> .\OctoSearch.exe search --regex connectionstring --output-file results.txt

To display all the variables in a html report we omit the regex to default to a greedy regex \w.. The html report has a client side search facility to filter variables for easier exploration.

> .\OctoSearch.exe search --output-file results.html --output-format html

This will give you a search UI that looks like:

If you would prefer it in Json:

> .\OctoSearch.exe search  --output-file results.json --output-format json

Once we have it in Json we can load and analyse it within powershell. For example to get all the variables marked as IsSensitive we could do:

> $variablesets = ConvertFrom-Json (gc -Raw .\results.json)
> ($variablesets | % { $_.Variables } | ? { $_.IsSensitive }).length
300

If you have any feedback or feature requests you can create an issue or reach me @naeemkhedarun.

Thanks to Aliostad for his analysis on the HttpClient and fschwiet for writing the .NET hosts library which I’ve now taken off my own to-do list.

There are two DNS level scenarios that I want to investigate:

• Connect to the production cloud service slot after a VIP swap. Clients connect to a cloud service using a cloudapp.net DNS name which points to the Azure Load Balancer distributing traffic over the nodes.
• Resolve the latest Traffic Manager configuration. Traffic manager is a DNS level load balancer. Clients resolve a traffic manager address to an IP address and connect directly to the endpoint. So what happens when the IP changes? We’ll try the scenario with both a Transient and Singleton HttpClient.

Transient

[Fact]
public async Task TransientTimeoutTest()
{
    HostsFile.Set(_hostname, "127.0.0.1");
    var client = new TransientHttpClient(new Uri($"http://{_hostname}:9200"));
    var timer = Stopwatch.StartNew();
    try
    {
        await new LoopUntilFailure(client).ExecuteAsync(
            () => Task.Run(() => HostsFile.Set(_hostname, "128.0.0.1")));
    }
    catch
    {
        _logger.WriteLine(timer.ElapsedMilliseconds.ToString());
    }
}

133721ms

The transient client eventually behaves as expected despite taking 133 seconds to respect the change. The ServicePointManager.DnsRefreshTimeout defaults to 120 seconds. This still leaves 13 seconds unaccounted for which I suspect is the final socket connection timeout.

A test isolating the connection to the non-responsive endpoint yields:

13086ms

I wasn’t able to find any configuration for this timeout within .NET but I didn’t manage to trace the framework source to an enumeration WSAETIMEDOUT. The timeout is controlled by the OS documented here.

TCP/IP adjusts the frequency of retransmissions over time. The delay between the first and second retransmission is three seconds. This delay doubles after each attempt. After the final attempt, TCP/IP waits for an interval equal to double the last delay, and then it closes the connection request.

You find the default values for your OS (in my case Windows Server 2016) by running:

#  netsh interface tcp show global
Querying active state...
TCP Global Parameters
----------------------------------------------
Initial RTO              : 3000
Max SYN Retransmissions  : 2

So the result should be (1 * 3000) + (2 * 3000) = 12000ms which explains the extra time. Now the result is understood, let’s re-run the test after dropping the DNS refresh timeout to 10 seconds.

ServicePointManager.DnsRefreshTimeout = 10000;

25026ms

So with a transient HttpClient a working way to stay up to date with traffic manager configuration is to tune the DnsRefreshTimeout property to a good value for your application.

Singleton

Using a singleton client will reuse the connection for many requests to reduce the overhead with starting new TCP connections. In this setup we still want the connection to be recreated occasionally so we get the latest DNS configuration.

[Fact]
public async Task SingletonDnsChangeTest()
{
    HostsFile.Set(_hostname, "127.0.0.1");
    var client = new SingletonHttpClient(new Uri($"http://{_hostname}:9200"));
    var timer = Stopwatch.StartNew();
    try
    {
        await new LoopUntilFailure(client).ExecuteAsync(
            () => Task.Run(() => HostsFile.Set(_hostname, "128.0.0.1")));
    }
    catch
    {
        _logger.WriteLine(
            $"{typeof(IHttpClient).Name} - {timer.ElapsedMilliseconds}");
    }
}

Cancelled after 180000

With a singleton HttpClient the connection is kept alive by default. This can be undesirable in configuration changes or scale out scenarios where you want your clients to connect to and use the new resources. Let’s try the DnsRefreshTimeout.

ServicePointManager.DnsRefreshTimeout = 10000;

Cancelled after 180000

Since the connection is open and kept open, we need to find a way to close it. There is another setting which controls the length of time a connection is held open for called ServicePointManager.ConnectionLeaseTimeout.

ServicePointManager.FindServicePoint(new Uri("http://testdns:9200"))
    .ConnectionLeaseTimeout = 10000;

145558

Unfortunately, having this setting alone isn’t enough based on our previous transient experiments; the DNS is still cached. Let’s combine the two settings.

33223

So now, despite using a singleton pattern within the code, our connections are being recreated and re-resolved up to every 20 seconds (both timeouts combined).

Is there a difference in latency and bandwidth when they are both hosted in the same region?

Test setup

To reflect the setup we have for our application, two VMs were provisioned in North Europe.

Source

• A3 (Large) Windows Cloud service
• Classic Virtual Network

Destination

• DS13v2 Linux Virtual machine
• ARM Virtual Network peered to the Classic VNet

Traceroute

I first wanted to test the latency and number of hops between the VMs. ICMP is not available for this test as we are hitting a public IP, however we can use TCP by using nmap.

PS C:\Users\user> nmap -sS -Pn -p 80 --traceroute 13.xx.xx.xx
HOP RTT     ADDRESS
1   ... 7
8   0.00 ms 13.xx.xx.xx

PS C:\Users\user> nmap -sS -Pn -p 80 --traceroute 10.xx.xx.xx
HOP RTT     ADDRESS
1   0.00 ms 10.xx.xx.xx

We can see that there are 8 hops over the public IP, and as we expect only a single hop over the peered network. Both routes are still extremely fast with negligible ping times. This confirms my collegues suspicions; despite connecting to a public address the traffic probably never leaves the datacenters perimeter network.

Bandwidth (iperf3)

To measure the bandwidth available between the VMs I’m using iperf3 which is cross platform. The test is run from the windows machine as a client and flows to the iperf server hosted on the linux box.

# Public IP test
.\iperf3.exe -c 13.xx.xx.xx -i 1 -t 30
# Peered network test
.\iperf3.exe -c 10.xx.xx.xx -i 1 -t 30

Surprisingly, both achieve the desired bandwith (1Gbps) for the selected VM sizes.

I was still curious if the performance profile was the same when upgrading both VMs to support 10Gbps networking. For this test both machines were upgraded to the DS14v2 VM size. To maximise the bandwidth I used iperfs -P switch to run concurrent workers. The buffer size was also increased to see the effect it has on the bandwidth.

#Public IP with 4 processes
.\iperf3.exe -c 13.7xx.xx.xx -i 1 -t 30 -P 4
#Peered network with 4 processes
.\iperf3.exe -c 10.xx.xx.xx -i 1 -t 30 -P 4
#Public IP with 4 processes and 32MB buffer
.\iperf3.exe -c 13.xx.xx.xx -i 1 -t 30 -P 4 -w 32MB
#Peered network with 4 processes and 32MB buffer
.\iperf3.exe -c 10.xx.xx.xx -i 1 -t 30 -P 4 -w 32MB

As expected, with the default values the peered network performed better although the difference was marginal. More surprisingly, the public network had a high thoroughput when the buffer size was increased and despite running the test multiple times I am unable to explain why.

For our workload and use case, I can say the performance difference between the two approaches is irrelevant. If you are evaluating whether you might gain network performance by switching to peered networking then I hope these numbers can help guide you. I would recommend running a similar test if you are choosing different VM sizes or workload.

Thankfully Chrome can help with that. Open the fabric explorer:

http://localhost:19080/explorer

Create the desktop app:

Then you can pin it to the taskbar as you would normally. You’ll get a window with all the extra Chrome removed.

Don’t forget to take advantage of the default windows taskbar keyboard shortcuts. I have it
pinned as the fourth taskbar item, so its quick to switch using win+4.

git config –global core.editor “‘C:\Program Files (x86)\Microsoft VS Code\code.exe’ -w”

And now when you run your git rebase -i the todo log will open in vscode.

• Unexpected actor bottleneck, they are single threaded.
• Time spent waiting on a bottlenecked actor.
• Large state affecting IO performance.

Reliable actors do not yet have interception interfaces to add in this kind of detailed telemetry, but with careful code its possible to do this with a dynamic proxy. I chose to use LightInject for this but most of the framework would do the same job. I use statsd and graphite as my telemetry platform and I’ve had good experiences with this nuget package

We need to intercept object on both sides of the network boundary to cover these scenarios.

• Service fabrics initialisation to trace OnLoadStateAsync, OnActivateAsync, etc…
• Fabric client initialisation to trace client interface calls IActorInterface.YourDoWorkMethodAsync.

We can trace the former by using service fabrics dependency injection support to initialise the actors with a proxy inbetween. First we override fabrics initialisation to use our DI container which has dynamic proxy support.

fabricRuntime.RegisterActorFactory(() => 
    ServiceLocator.Current.GetInstance<YourActor>());

Next we tell our DI container to resolve these types with a proxy that includes our telemetry interceptor.

serviceRegistry.Intercept(registration => 
    typeof(IActor).IsAssignableFrom(registration.ServiceType),
    (factory, definition) =>
    {
        definition.Implement(factory.GetInstance<ActorInterceptor>);
    });

This will catch the timings for any calls to actors made by the fabric system. Now we need to get the timings for all the calls we make, both actor to actor and client to actor.

public class ActorFactory : IActorFactory
{
    readonly ActorInterceptor _actorInterceptor;
    readonly ConcurrentDictionary<Type, Type> _proxies = 
        new ConcurrentDictionary<Type, Type>();
    public ActorFactory(ActorInterceptor actorInterceptor)
    {
        _actorInterceptor = actorInterceptor;
    }
    public T Get<T>(ActorId actorId) where T : IActor
    {
        var proxyType = _proxies.GetOrAdd(typeof(T), type =>
            new ProxyBuilder().GetProxyType(
                new ProxyDefinition(typeof(T), true).Implement(
                    () => _actorInterceptor)));
        return (T)Activator.CreateInstance(proxyType, 
            new Lazy<T>(() => ActorProxy.Create<T>(actorId)));
    }
}

Above we’ve created a factory class which should be used by clients and actors to create the proxied ActorProxies. We cache the generated proxy types in a thread safe dictionary as they are expensive to create.

Lastly we need the intercetor itself. We need to be sympathetic towards:

• All actor calls return a Task.
• Avoid blocking calls by calling Result or Wait on the task.

We can use a task continuation to handle the writing of telemetry together with a closure to capture the timer. If there is a return value we should return it, and for whatever reason that value is not a Task then we won’t try to add the continuation.

public class ActorInterceptor : IInterceptor
{
    public Object Invoke(IInvocationInfo invocationInfo)
    {
        var name = string.Format("Actors.{0}.{1}",
            invocationInfo.Proxy.Target.GetType().Name,
            invocationInfo.Method.Name);
        var timer = Metrics.StartTimer(name);
        var returnValue = invocationInfo.Proceed();
        (returnValue as Task)?.ContinueWith(task =>
        {
           timer.Dispose();
        });
        return returnValue;
    }
}

If you have your metrics library configured to push to a graphite backend you can use the following query to graph it:

stats.timers.actors.*.*.mean
stats.timers.actors.*.*.count

We’ll take a look at:

• Castle Dynamic Proxy - This is likely the most popular dynamic proxy library for .NET.
• LightInject Interceptor - I hadn’t heard of this one until I started using LightInject.
• NProxy - Also a new one to me, but I’ve added it since it is easy to use.

Creating proxy instances

I was quite surprised to see such a difference between the frameworks. I guessed that both NProxy and Castle cache their proxy types internally, which LightInject expects you to handle your own caching. Something good to bear in mind!

After caching the proxy type things are a little more competitive:

I still think the code can be more optimal in all cases, so I reduced everything as much as possible to a single call to activate the proxy type. I’ve included timings for Activator.CreateInstance and the standard constructor against the non-proxy type as a baseline.

Things are much closer now! The difference between Castle and LightInject are negligible. There might be a way to optimise NProxy further but the API didn’t yield any obvious optimisations.

Calling proxied methods

Now let’s take a look at the runtime overhead of calling a proxied object. I’ve included an unproxied instance as a baseline.

Surprisingly there is no overhead with any of the libraries with calling the proxied object. The graph looks skewed due to how close the results are and the timings are in nanoseconds. This is great news and we can use whichever library we want guilt-free.

You can review the code for the benchmarks on github.

var fabricClient = new FabricClient();
var partitions = await fabricClient.QueryManager.GetPartitionListAsync(
    new Uri($"fabric:/App/{serviceName}"));
var partition = partitions[0] //Select the first partition;
var partitionLowKey = ((Int64RangePartitionInformation)partition.PartitionInformation).LowKey;

The result of GetPartitionListAsync should never change for a service as you can’t change the partition information after a service has been created. It would be safe and give better performance to cache this.

The endpoint of the primary replica however can move between machines, so this does need to be resolved more frequently. You can also cache this if you have a retry strategy that will re-resolve after an EndpointNotFoundException.

var resolver = new ServicePartitionResolver(() => fabricClient);
var resolvedPartition = await resolver.ResolveAsync(new Uri($"fabric:/App/{serviceName}"), partitionLowKey, CancellationToken.None);

• Named Partition - This is a fixed name for each partition configured at deploy time.
• Ranged Partition - This uses an Int64 range to decide which partition a numbered key falls in.

More information can be found here.

Named Partitioning

A named partition allows you to specify explicitly which partition you want to access at runtime. A common example is to specify A-Z named partitions and use the first letter of your data as the key. This splits your data into 26 partitions.

<Service Name="TestService">
  <StatefulService ServiceTypeName="TestServiceType" 
                   TargetReplicaSetSize="3" 
                   MinReplicaSetSize="2">
    <NamedPartition>
      <Partition Name="a"/>
      <Partition Name="b"/>
      ...
      <Partition Name="z"/>
    </NamedPartition>
  </StatefulService>
</Service>

The advantages to this are that it is simple and you know which partition your data goes in without a lookup. Unfortunately as we will test later, you are unlikely to get a good distribution of your data across the partitions.

Ranged Partitioning

With a ranged partition the fabric tooling by default uses the entire Int64 range as keys to decide which partition. It will then convert these into ranges or buckets depending on the partition count.

<Service Name="TestService">
  <StatefulService ServiceTypeName="TestServiceType" 
                   TargetReplicaSetSize="3" 
                   MinReplicaSetSize="2">
    <UniformInt64Partition PartitionCount="26"
                           LowKey="-9223372036854775808" 
                           HighKey="9223372036854775807" />
  </StatefulService>
</Service>

However to be able to lookup a partition we need a function which can reduce our data to an integer value. To use the configuration above we can convert our strings into an Int64.

var md5 = MD5.Create();
var value = md5.ComputeHash(Encoding.ASCII.GetBytes(value));
var key = BitConverter.ToInt64(value, 0);
var client = ServiceProxy.Create<ITestService>(
                    key, 
                    new Uri("fabric:/App/TestService"))

1. Hash the value to a fixed length byte array.
2. Convert the array to an Int64.
3. Create the client with the calculated key to connect to the service on that partition.

Ranged Partition with Consistent Hashing

Rather than use the ranges, you can fix your keys and plug in your own hash algorithm to resolve the partition.

<Service Name="TestService">
  <StatefulService ServiceTypeName="TestServiceType" 
                   TargetReplicaSetSize="3" 
                   MinReplicaSetSize="2">
    <UniformInt64Partition PartitionCount="26" 
                           LowKey="0" 
                           HighKey="25" />
  </StatefulService>
</Service>

We now have a key range limited to 0-25 rather than the entire Int64 range. We can resolve a client connected to this partition in the same way, however this time we need to compute a key that fits in this smaller range. I’m using the jump consistent hash implementation in hydra.

var shard = new JumpSharding().GetShard(value, 26);
var client = ServiceProxy.Create<ITestService>(
                    shard, 
                    new Uri("fabric:/App/TestService"))

1. Call get shard with the value and number of partitions to distribute across.
2. Create the client with the calculated key to connect to the service on that partition.

Distribution

To benchmark the distribution we have a list of around 17000 real email addresses. This should give us an idea of how the sharding strategies will distribute the data across 26 partitions. Another advantage of using one of the Int64 methods is that they can be used with any amount of partitions.

We are looking for an even number of accounts allocated to each partition.

We can see from those results that sharding using the first character of an email produces wildly different partition sizes, not what we want! Both the jump hash and integer ranging methods produced very even parition sizes.

Conclusion

Based on these results I would use the ranged partitioning method, it produces provides good balancing and is fast to compute. An additional advantage is you do not need to know the partition count in the code, just map your data to an Int64 and service fabric will do the rest.

• Install a logging framework like NLog or log4net.
• Put try catches around your main program logic and include contextual information with the exception.
• If its a web application use a logging middleware for your web framework.

However this will not log unhandled exceptions from places you couldn’t forsee. So let’s log these just in case anything goes wrong.

Any exceptions which crash the application can be handled using the UnhandledException event.

AppDomain.CurrentDomain.UnhandledException += (sender, args) =>
{
    Log.Exception((Exception)args.ExceptionObject);
};

This will help you diagnose fatal errors. Unfortunately not all exceptions are fatal, and if you have any timers, unawaited async or unhandled task pool exceptions these can cause your application to behave unexpectedly without you knowing about it.

You can use the UnobservedTaskException to catch some of those ones:

TaskScheduler.UnobservedTaskException += (sender, args) =>
{
    if (!args.Observed)
    {
        Log.Exception(args.Exception);
    }
};

If you know any other events or ways to get more of these unexpected errors please let me know!

Attach a disk

To attach a data disk to the virtual machine you can use the following configuration just below the osDisk definition.

{
      "dataDisks":[{
            "lun":0,
            "name":"datadisk",
            "diskSizeGB":1000,
            "createOption":"Empty",
            "vhd":{
                  "Uri":"[variables('dataDiskUri')]"
            }
      }]
}

• lun (Logical Unit Number) should be unique for each attached disk and starts from zero.
• diskSizeGB has a maximum of 1000 (1 TB) in Azure currently.
• createOption should be “Empty” for a new disk.
• vhd.Uri is the full path you want the blob created in.

If we deployed the template with this now, we will get a raw unformatted disk with no drive label. You can view it in Disk Manager or by running the following PowerShell on the box.

Get-Disk | Where partitionstyle -eq 'raw'

We need one more resource to complete the deployment and give us a formatted disk. We can use the CustomScriptExtension to run a PowerShell script which will format any raw attached disks (in case you want to do multiple drives).

{
   "type":"Microsoft.Compute/virtualMachines/extensions",
   "name":"[concat(variables('vmName'),'/InitialiseDisks')]",
   "apiVersion":"2015-05-01-preview",
   "location":"[resourceGroup().location]",
   "dependsOn":[
      "[concat('Microsoft.Compute/virtualMachines/', variables('vmName'))]"
   ],
   "properties":{
      "publisher":"Microsoft.Compute",
      "type":"CustomScriptExtension",
      "typeHandlerVersion":"1.2",
      "settings":{
         "fileUris":[
            "[variables('initialiseDisksScript')]"
         ],
         "commandToExecute":"[concat('powershell -ExecutionPolicy Unrestricted -file ',parameters('scriptName'))]"
      }
   }
}

• dependsOn will need to be the resource path for our virtual machine to ensure this extension is run after its provisioned.
• fileUris is an array of any files we want downloaded and made available to execute. This needs to be a path to an azure blob, unfortunately you cannot link to any other domains like github.
• commandToExecute is a cmd.exe command, so we need to call out to powershell to invoke our script. The working directory will have your downloaded file already there, so you don’t need the full path to execute it, just the name.

You will need to upload the following script to a blob:

Get-Disk | ` 
Where partitionstyle -eq 'raw' | ` 
Initialize-Disk -PartitionStyle MBR -PassThru | ` 
New-Partition -AssignDriveLetter -UseMaximumSize | ` 
Format-Volume -FileSystem NTFS -NewFileSystemLabel "datadisk" -Confirm:$false