coffeewheat

[coffeewheat]

kubectl helper

https://github.com/kuritka/kubernetes_helper/blob/master/network-utils.yaml

[coffeewheat]

connecting to SFTP / SSH

~/.ssh/ needs to be setted to 0700 authorization_keys needs to be setted to 0600 sftp -oIdentityFile=pk.key edt2@localhost ssh -i ./pk.key edt2@localhost good luck! :)

[coffeewheat]

How to store private key in environment variable

This is necessary for keeping keys in environment variables. It does work under Linux but on Windows you cannot make it. # store IFS first SAVEIFS=$IFS IFS='' #udate environment variable export EDT_SFTP_PRIVATE_KEY="$(cat Edt_Private.key)" #check whether it is well exported echo $EDT_SFTP_PRIVATE_KEY #set IFS back IFS=$SAVEIFS Another solution is to keep key stored in base64 format so you doesnt handle enters: export EDT_SFTP_PRIVATE_KEY="$(cat Edt_Private.key) | base64 -w0"

[coffeewheat]

Database performance

I created very simple M:N relation between batches and ISISNs. Each Batch could contain up to the 1000 ISISNs.

Here are some statistics:

Very fisrt run

SQL Server parse and compile time:   CPU time = 0 ms, elapsed time = 6 ms.(2 rows affected) SQL Server Execution Times:   CPU time = 10077 ms,  elapsed time = 4012 ms. SQL Server parse and compile time:   CPU time = 0 ms, elapsed time = 0 ms.

Next run

SQL Server parse and compile time:   CPU time = 0 ms, elapsed time = 0 ms.(2 rows affected) SQL Server Execution Times:   CPU time = 8921 ms,  elapsed time = 2428 ms. SQL Server parse and compile time:   CPU time = 0 ms, elapsed time = 0 ms.

Join with four batches

[coffeewheat]

How to update .bashrc

echo 'export PATH="$PATH:/opt/mssql-tools/bin"' >> ~/.bash_profile echo 'export PATH="$PATH:/opt/mssql-tools/bin"' >> ~/.bashrc source ~/.bashrc

[coffeewheat]

Connecting to github from Fedora

Here is another way how to connect to github from Fedora:

git config --global url."git@github.com:".insteadOf "https://github.com/" g git config --global user.email="your@email.com" git config --global user.name="Your Name" cd ~/.ssh/ # generate key pair and add public key to your github $ ssh-keygen -t rsa -C "your@email.com" #add generated public key to your GIT. #https://github.com/settings/keys ssh -vt <user>@github.com # When it prompts error ... #Are you sure you want to continue connecting (yes/no)? # Say 'yes' #Now your remote host is added to the known_hosts

git clone https://github.com/XYZ/edt-platform go get -v ./...

[coffeewheat]

How to store environment variable with new lines

working well, when you need to store private keys into environment variables:

$ SAVEIFS=$IFS $ IFS='' $ export EDT_SFTP_PRIVATE_KEY="$(cat ~/key.private)" $ echo $EDT_SFTP_PRIVATE_KEY $ IFS=$SAVEIFS

more info about $IFS: https://unix.stackexchange.com/questions/164508/why-do-newline-characters-get-lost-when-using-command-substitution

[coffeewheat]

Acessing RDP on AZURE RHEL 7.4

Here is step by step what needs to be done on VM to. First you need to install GNOME on your machine and thanallow remote desktop to access from Windows.

`#systemctl stop firewalld` `#systemctl disable firewalld` `#yum install   https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm` `#yum install xrdp tigervnc-server xterm` `#systemctl enable xrdp` `#systemctl enable xrdp-sesman` `sudo systemctl start xrdp` `sudo systemctl start xrdp-sesman` #Private key authentication is forbidden for RDP, so you need setup #password for user `sudo passwd <your user>` #following commands must show green result `sudo systemctl status xrdp` `sudo systemctl status xrdp-sesman` #Going with disabled firewall is antipattern # Please spend some time by investigating correct # iptables setup and enable firewall back. #After running GoLand installation you may see blank dialog. #This would be fixed by going to `/etc/xrdp/xrdp.ini` #and setting to `max_bpp=24`. #Restart RDP than..

[coffeewheat]

Setup GIT for SSH on Windows 10

Once we switched to go we started to clone project including all dependencies 

i.e. 

go get github.com/Deutsche-Boerse/gRPC

normally we run :

git config --global url."git@github.com:".insteadOf "https://github.com/"

go get github.com/Deutsche-Boerse/gRPC - Clone project

go get ./... - Install project dependencies

Because traffic goes through SSH you need to make additional steps on your machine:

git config --global user.email="your@email.com" git config --global user.name="Your Name" #open Git Bash from START menu and run $ ssh-keygen -t rsa -C "your@email.com" # generate key pair and add public key to your github (https://github.com/settings/keys) #add github public key to your GIT $ ssh-keyscan -t rsa github.com >> ~/.ssh/known_hosts

And that's all! enjoy...

[coffeewheat]

View Kubernetes Cluster on Azure

To see kubernetes cluster you need to install kubernetes and azure-cli. Than run following commands 

az login az aks get-credentials "resource-group EDT4AKS "name edtAKSCluster kubectl config use-context edtAKSCluster # switch kubectl context to edtAKSCluster kubectl get services

dt-api-gateway-deployment LoadBalancer 10.0.229.117 40.114.248.163 80:32367/TCP 38d edt-api-gateway-service NodePort 10.0.127.236 <none> 80:30811/TCP 38d edt-etl-service ClusterIP 10.0.18.93 <none> 9020/TCP 38d edt-globalnotes-service ClusterIP 10.0.217.54 <none> 9030/TCP 38d edt-interaction-service ClusterIP 10.0.202.8 <none> 9010/TCP 38d edt-upload-service ClusterIP 10.0.52.91 <none> 9000/TCP 38d edt-validator-service ClusterIP 10.0.105.41 <none> 9040/TCP 38d edt-workflow-service ClusterIP 10.0.6.163 <none> 8010/TCP 38d kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 38d

[coffeewheat]

Publish/Subscribe vs Producer/Consumer

Although both Publisher/Subscriber and Producer/Consumer terms are related to messaging, they are different and can't be used interchangeably.

Publisher/Subscriber is a messaging pattern where a publication is distributed to multiple receivers. Whereas in Producer/Consumer pattern, Producer is the sender of messages and consumer is the receiver of messages. Producer and consumers are an integral part of both Publisher/Subscriber and Point-to-Point messaging patterns.

more info about produce consumer is here

[coffeewheat]

Covariance x Contravariance

This is one of the moost confusing generic parts of C# (exists from 4.0)

The goal is to moving from specific to general.  Look at the code below: thanks to out T  we are able to do something like this:

public interface IFoo<out T> { T GiveMeFoo(); } public class Program { public static void Main(string[] args) { IFoo<string> foos = null; IFoo<object> generals = foos; /*covariance enables such analogy with generic types*/ string foo = "text"; object general = foo; } }

If you move “out” from interface you cannot convert from IFoo<string> to IFoo<object>. 

Now something more difficult:

public interface IFoo<out T> { T GiveMeFoo(); } public interface IBar<T> : IFoo<T> { } public class Program { public static void Main(string[] args) { IBar<string> barStr = null; IBar<object> barObj = barStr; /*BUILD FAILS because IBar has not covariant T*/ IFoo<object> fooObj = barStr; /*is Ok, because IBar derives from IFoo and IFoo has out T*/ } }

Invariant IList<T>

Problem with IList is similiat to IBar<T> in previous example.  It is invariant (it doesnt allow either covariance or contravariance ). Take look at the next example : 

public interface IFoo<out T> { IEnumerable<T> GetAllFoos(); /*WORKING*/ IList<T> GetSomeFoos(); /*DOESNT COMPILE*/ } public interface IEdible { }; public abstract class Fruit : IEdible { } public class Apple : Fruit{ } public class Banana: Fruit { } public class Pizza : IEdible { } public class Program { public static void Main(string[] args) { List<Banana> bananas = new List<Banana>(); List<IEdible> edibles = bananas; /*DOESNT COMPILE, IList doesnt allow Covariance*/ IEnumerable<IEdible> eedibles = bananas; /* working IEnumerable allows Covariance*/ IEnumerable<Fruit> fruit = bananas; /* working */ edibles.Add(new Banana()); EatAll(bananas); /*DOESNT COMPILE*/ } public static void EatAll(List<IEdible> edibles) { foreach (var item in edibles) { //Do stuff here } } }

 Don’t worry, compiler always keep you safe and tell’s you when covariently or contravariently you shouldn’t be. 

Covariance is no type conversion it is allows Reference conversion (you still working with the same object but in the different kind of way).

Contravariance

Because Covariance allows only returned values you cannot mix it with input values. That’s why contravariance exists.

interface IFoo<out T, in U> where T : string, where U : string 

{

      T QuerySomething(U param);

}

[coffeewheat]

SQL Update script

/* sets NotionalCurrency to Deliverablecurrency */ SET XACT_ABORT ON SET NOCOUNT ON PRINT N'Preparing data.' declare @transactionsToUpdate table ( DeliverableCurrency1ExtRefInID int, DeliverableCurrency2ExtRefInID int, NotionalCurrency1ExtRefInID int, NotionalCurrency2ExtRefInID int) BEGIN PRINT 'Starting..' DECLARE @cnt INT SELECT @cnt = COUNT(*) FROM @transactionsToUpdate BEGIN TRANSACTION /* UPDATE CODE HERE. Don't forget to fill AuditLog */ IF(XACT_STATE() = -1) BEGIN ROLLBACK TRAN PRINT N'ROLLED BACK' END ELSE BEGIN IF(@@TRANCOUNT = 0 ) BEGIN PRINT 'ERROR OCCURED' END ELSE BEGIN ROLLBACK TRAN PRINT N'HA HA HA! transaction was ROLLBACKED because COMMIT is commented!' --PRINT N'COMMITED' --COMMIT TRAN END END END

[coffeewheat]

Horizontal VS Vertical Deployment

More info here

[coffeewheat]

Compression ratio 1:454

Our beamer project will compress huge XML and chunks them into small zipped files. We started with simulation 3GB input XML (top limit) . This valid xml is composed from many repeating parts differ by its growing ID. Because smaller parts are repeating in less than one minute we were able to compress 3GB data into one 6MB zipped chunk !

Beautiful , int’t it ?

[coffeewheat]

Docker + Ansible + Jenkins

We are looking for right build pipe of our new .net core project - BEAMER. I found this article that brings me a light into chaos 

Basic (Continuous Integration?) Deployment with Ansible, Docker, Jenkins and Git

Jenkins manages WHEN and WHAT to do.

Ansible manages HOW and WHERE to do.

Docker ISOLATE your application with the host, if the host can run docker, you can deploy your application.