<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Talking Tech Around - Medium]]></title>
        <description><![CDATA[Talking about startups and diffrent types of technologies that interests me. Most of the technologies will be in the new type of information technology for ex. Cloud Services , SaaS , PaaS , IaaS , Containers , Machine Learning &amp; AI in IT and of course how IoT will be there also. - Medium]]></description>
        <link>https://medium.com/talking-tech-all-around?source=rss----2dd325eb3f7a---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Talking Tech Around - Medium</title>
            <link>https://medium.com/talking-tech-all-around?source=rss----2dd325eb3f7a---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 12:25:29 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/talking-tech-all-around" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[VMware Cloud Native Master Specialist (5V0–71.19) — Exam Experience and Tips]]></title>
            <link>https://medium.com/talking-tech-all-around/vmware-cloud-native-master-specialist-5v0-71-19-exam-experience-and-tips-99e4d52deb12?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/99e4d52deb12</guid>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[cloud-native]]></category>
            <category><![CDATA[containers]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[vmware]]></category>
            <dc:creator><![CDATA[Itay Talmi]]></dc:creator>
            <pubDate>Sun, 29 Nov 2020 08:42:45 GMT</pubDate>
            <atom:updated>2020-11-29T08:42:45.542Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>VMware Cloud Native Master Specialist (5V0–71.19) — </strong>Exam Experience and Tips</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/340/1*VFi7tUKxVYHHv4tbvCxbyw.png" /></figure><p>I recently passed my VMware Cloud Native Master Specialist exam.</p><p>Although the main purpose of this was to earn the Cloud Native Master Services Competency (MSC) for our company (more on that <a href="https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/product/vmware-MSC-FAQ-0420.pdf">here</a>), it was an interesting experience which led me to explore several concepts and open source projects/technologies I was not familiar with.</p><p>I would say this exam is a tough one. However, most of the questions are actually based on real-world scenarios, which is nice, so if you have worked with Kubernetes-based platforms and their complementary technologies, you should be fine. Also, since the CKA certification is a prerequisite for this exam, you should already be familiar with a lot of the concepts covered by the exam.</p><p>Since there is not much information available on this exam out there, I thought it would be helpful to share my experience and tips.</p><p><strong>Exam Information</strong></p><p>The Cloud Native Master Specialist exam (5V0–71.19) consists of 67 questions in a single/multiple choice format. You have 100 minutes to complete and the passing score is 300 (60%). It is a proctored exam and can either be taken remotely or at a test center. I personally chose the online option. The registration is handled by Pearson VUE, and if you go for the online option, you will have to use their secure browser application in order to complete the onboarding process and access the exam environment on the day of the exam.</p><p>Once you complete the exam, you will immediately receive the results and your score.</p><p><strong>Prerequisites</strong></p><p>The only prerequisite for this exam is the <a href="https://www.cncf.io/certification/cka/">CKA (Certified Kubernetes Administrator)</a> certification. You are required to submit proof that you hold an active CKA certification from the CNCF. This is done via the VMware Certification Manager portal. Once your CKA certification is confirmed by VMware, you can complete the registration process and schedule your exam.</p><p><strong>Required Knowledge and Useful Resources</strong></p><p>First of all, the exam has absolutely nothing to do with any traditional VMware products, nor with PKS/TKG. I think this is an important clarification. It is completely focused on Kubernetes and complementary technologies, and since the CKA certification is a prerequisite for it, by the time you take the exam, you should already be familiar with a lot of the concepts. If you are not already CKA certified and planning to get started, I strongly recommend completing <a href="https://www.udemy.com/course/certified-kubernetes-administrator-with-practice-tests/">Mumshad Mannambeth’s CKA course on Udemy</a>, which is, by far, one of the best resources available for the CKA exam.</p><ul><li>The starting point, just like any other VMware exam, should be the <a href="https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/certification/vmw-ms-cloud-native-2019.pdf">official exam blueprint</a>. It is absolutely crucial to go through all of the links provided in the blueprint.</li><li>I also strongly recommend completing the courses provided by <a href="https://kube.academy/">KubeAcademy</a>.</li><li>It is important to understand security concepts, such as secrets, pod security policies, network policies, RBAC and the different authentication methods as well as complementary technologies that can be leveraged for authentication, audit policies, admission control, the integration between open policy agent and Kubernetes, etc.</li><li>Understand the integration between Kubernetes and Prometheus, Fluent Bit and log aggregation methods in general.</li><li>Understand Readiness and Liveness probes, as well as the differences between the two.</li><li>Be familiar with Kubernetes networking, service types, as well as the different methods that can be used to expose applications externally, such as LoadBalancer/Ingress. Also, basic AWS knowledge could help.</li><li>Be familiar with complementary technologies provided by VMware/Heptio, such as Velero, Sonobuoy, etc.</li><li>Be familiar with Dockerfile best practices/troubleshooting, building containers and leveraging multistage builds.</li><li>Be familiar with Helm best practices in general.</li><li>Understand the 12-factor app concepts — this is very important for the exam and in general!</li><li>Be familiar with the popular CNIs and their differences.</li></ul><p><strong>Conclusion</strong></p><p>If you thoroughly go through the exam blueprint and the resources/topics I mentioned above, you should be able to pass the exam.</p><p>Although some of the questions on the exam may seem overwhelming at first — don’t panic, flag questions you don’t know how to answer and go back to those questions at the end of the exam. Since you have about 1.5 minutes per question, you have to carefully manage your time during the exam.</p><p>Good luck!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=99e4d52deb12" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/vmware-cloud-native-master-specialist-5v0-71-19-exam-experience-and-tips-99e4d52deb12">VMware Cloud Native Master Specialist (5V0–71.19) — Exam Experience and Tips</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[KubeCon 2019 Barcelona — Impressions and More…]]></title>
            <link>https://medium.com/talking-tech-all-around/kubecon-2019-barcelona-impressions-and-more-492ce80ee178?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/492ce80ee178</guid>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[containers]]></category>
            <category><![CDATA[kubernetes]]></category>
            <dc:creator><![CDATA[Etgar Fishel]]></dc:creator>
            <pubDate>Thu, 04 Jul 2019 07:49:34 GMT</pubDate>
            <atom:updated>2019-07-04T07:49:34.197Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/703/1*g7239Z_1DAx2l9SpFH2nag.png" /></figure><h3><strong>KubeCon 2019 Barcelona — Impressions and More…</strong></h3><p>This year, I was fortunate enough to attend #KubeCon and #CloudNativeCon Europe, in Barcelona. #KubeCon 2019 was huge, with lots of great sessions, and I’d like to outline key points which I find as the most important and relevant from #DevOps perspective.</p><p>For those who ask what KubeCon is, KubecCon Is a #Kubernetes Conference, an event dedicated to the mostly used container orchestrator: Kubernetes and all surrounding ecosystem and community. Kubernetes is an open-source container-orchestration system for automating applications,deployment, scaling, and management. Its original design based on Google’s Borg and is now maintained by the Cloud Native Computing Foundation (CNCF) Ecosystem.</p><p>First, we need to understand that Kubernetes is not one stop shop software as one may think; but it’s a framework with a vast ecosystem and you should build your puzzle according to the business needs of your organization and your specific use cases. In this my article, I would like to take you to a quick tour, introducing you with few relevant tools and platforms so you will be bit more familiar with some of your options as I leaned in the conference.</p><p><strong>Monitoring</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/142/1*XaJegaDIu2HocNnfKIl3QA.png" /></figure><p>● <a href="https://prometheus.io/"><strong>Prometheus</strong> </a>— Prometheus is an open-source system monitoring and alerting toolkit originally built at SoundCloud. Since its inception in 2012, many companies and organizations have adopted Prometheus, and the project has a very active developer and user community. It is now a standalone open-source project and maintained independently of any company. To emphasize this, and to clarify the project’s governance structure, Prometheus joined the Cloud Native Computing Foundation in 2016 as the second hosted project, after Kubernetes.</p><p><strong>Logging</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/129/1*GqevkbdtpVkKW7GDN9K49Q.png" /></figure><p>● <a href="https://www.fluentd.org/"><strong>Fluentd</strong> </a>— decouples data sources from backend systems by providing a unified logging layer in between. This layer allows developers and data analysts to utilize many types of logs as they are generated. Just as importantly, it mitigates the risk of “bad data” slowing down and misinforming your organization. A unified logging layer lets you and your organization make better use of data and iterate more quickly on your software.</p><p><strong>Tracing:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/118/1*qQQ0u1ZGpg9pPPEhvrR4Kg.png" /></figure><p>● <a href="https://www.jaegertracing.io/"><strong>Jaeger</strong> </a>— inspired by Dapper and OpenZipkin, is a distributed tracing system released as open source by Uber Technologies. It is used for monitoring and troubleshooting microservices-based distributed systems, including:</p><p>● Distributed context propagation</p><p>● Distributed transaction monitoring</p><p>● Root cause analysis</p><p>● Service dependency analysis</p><p>● Performance / latency optimization</p><p>● Uber published a blog post, Evolving Distributed Tracing at Uber, where they explain the history and reasons for the architectural choices made in Jaeger</p><p><strong>Service mesh:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/126/1*6P_rCTt7YCeXHQe7l0PJ2Q.png" /></figure><p>● <a href="https://istio.io/"><strong>Istio</strong> </a>— is a service mesh for Kubernetes and other frameworks. Istio lets you connect, secure, control, and observe services. At a high level, Istio helps reduce the complexity of these deployments, and eases the strain on your development teams. It is a completely open source service mesh that layers transparently onto existing distributed applications.</p><p><strong>Visualization UI</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/264/1*bUochZRS_yROHJOvFzzyEQ.png" /></figure><p>● <a href="https://grafana.com/"><strong>Grafana</strong> </a>— Grafana is an open source visualization tool that can be used on top of a variety of different data stores but is most commonly used together with Prometheus and Elasticsearch. Grafana allows you to query, visualize, alert on and understand your metrics no matter where they are stored. Create, explore, and share dashboards with your team and foster a data driven culture.</p><p><strong>Infrastructure provisioning</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/159/1*vyqEiiE-s3N2ngrMQ0ohcQ.png" /></figure><p>● <a href="https://www.terraform.io/"><strong>Terraform</strong> </a>— Terraform is an open-source infrastructure as code software tool created by HashiCorp. It enables users to define and provision a datacentre infrastructure using a high-level configuration language known as HashiCorp Configuration Language, or optionally JSON.</p><p><a href="https://www.skywiz.io/"><strong>SkyWiz</strong></a><strong>, By TeraSky</strong> Is official HashiCorp system Integrator</p><p><strong>Configuration management:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/109/1*QT7MyJjxcFy4sWX1A3J5Sg.png" /></figure><p>● <a href="https://helm.sh/"><strong>Helm</strong> </a>— uses a packaging format called charts. A chart is a collection of files that describe a related set of Kubernetes resources. A single chart might be used to deploy something simple, like a simple pod, or something complex, like a full web app stack with HTTP servers, databases, caches, and so on.</p><p><strong>Storage</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/94/1*u7Uo4bU2Q5DnaW93PN246w.png" /></figure><p>● <a href="https://rook.io/"><strong>Rook </strong></a><strong>— </strong>Rook turns storage software into self-managing, self-scaling, and self-healing storage services. It does this by automating deployment, bootstrapping, configuration, provisioning, scaling, upgrading, migration, disaster recovery, monitoring, and resource management. Rook uses the facilities provided by the underlying cloud-native container management, scheduling and orchestration platform to perform its duties.</p><p><strong>Security</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/270/1*-PtUO59uIFb7vHZKBJTEVg.png" /></figure><p>● <a href="https://www.twistlock.com/"><strong>Twistlock </strong></a><strong>— </strong>Twistlock is cloud native security platform, providing holistic coverage across hosts, containers, and serverless computing in a single platform. Twistlock is cloud-native and API-enabled itself, protecting all your workloads throughout their lifecycle, regardless of what underlying compute technology powers them.</p><p>Twistlock’s cloud native cyber-security platform provides full visibility into container activity, allowing organizations to detect and prevent suspicious activity and attacks, providing transparent, automated security while helping to enforce policy and simplify regulatory compliance.</p><p><a href="https://www.skywiz.io/"><strong>SkyWiz</strong></a> by TeraSky Is <strong>Twistlock</strong> official partner</p><p>I hope that by this scanning and short introduction for these relevant tools and platforms I helped you to arrange your preferences and options when working in containers based environment. If you still have any issues, concerns or questions you can contact me directly or <a href="http://info@skywiz.io/">info@skywiz.io</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=492ce80ee178" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/kubecon-2019-barcelona-impressions-and-more-492ce80ee178">KubeCon 2019 Barcelona — Impressions and More…</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[PostgreSQL High Availability on K8s — The good hard way]]></title>
            <link>https://medium.com/talking-tech-all-around/postgresql-high-availability-on-k8s-the-good-hard-way-554072a2ba8e?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/554072a2ba8e</guid>
            <category><![CDATA[postgres]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[stolen]]></category>
            <category><![CDATA[helm]]></category>
            <dc:creator><![CDATA[Daniel Vaknin]]></dc:creator>
            <pubDate>Mon, 11 Mar 2019 08:52:41 GMT</pubDate>
            <atom:updated>2019-03-03T19:30:38.662Z</atom:updated>
            <content:encoded><![CDATA[<h3>PostgreSQL High Availability on K8s — The good hard way</h3><p><strong>You have an existing instance of PostgreSQL on Kubernetes and want to migrate it to a highly available cluster? Stolon to the rescue! But hold your excitement, initializing Stolon with existing postgres data is not that simple…</strong></p><p>So you’ve already deployed PostgreSQL in your K8s cluster as part of your application, but now you want to make it highly available, since, of course, you don’t want to loose any of your precious data.</p><p>There are several open-source projects which will help you achieve this goal, but in this post I’ll be focusing on <strong><em>Stolon</em></strong>, which, according to their documentation is:</p><blockquote>a cloud native PostgreSQL manager for PostgreSQL high availability. It’s cloud native because it’ll let you keep an high available PostgreSQL inside your containers (kubernetes integration) but also on every other kind of infrastructure (cloud IaaS, old style infrastructures etc…)</blockquote><p>You can read more about Stolon in their GitHub page: <a href="https://github.com/sorintlab/stolon">https://github.com/sorintlab/stolon</a></p><p>In our case, we’ve chosen to deploy stolon on our K8s cluster, using Helm.</p><p>The official Stolon helm chart can be found here: <a href="https://github.com/helm/charts/tree/master/stable/stolon">https://github.com/helm/charts/tree/master/stable/stolon</a></p><p>In addition, we’ve modified Stolon’s chart to deploy Stolon with PVCs (and PVs), so a few adjustments had to be made to the original chart.</p><h3>Stolon installation overview</h3><p>When deploying Stolon helm chart on your K8s cluster, it’ll create a new Keeper data directory with a few files necessary for the cluster to operate, where among those files, you’ll find the postgres folder with the same structure as the good old postgres folder in your original PostgreSQL deployment.</p><h3><strong>What is the problem?</strong></h3><p>As I’ve stated before, stolon will create a <strong>new </strong>postgres folder, with “empty” database. <strong>But how can we use our existing </strong><strong>postgres and existing data in a new deployment of Stolon?</strong></p><h3><strong>What is the solution?!</strong></h3><p>There are 2 ways to “migrate” data from PostgreSQL to Stolon:</p><ol><li>pg_dump and pg_restore — Although this is a quite simple solution for this task, not only that it requires the original PostgreSQL database to be available to perform the pg_dump, but you also won’t be able to write any new data to the database until the pg_restore will complete on the new Stolon cluster. I’ll not elaborate on this solution in this post, since it’s quite straight forward, but not suitable for every use case.</li><li>Start Stolon with existing postgres folder (containing all PGDATA from the original PostgreSQL database) — Although a bit complicated solution (as you’ll see in this post), it’s quite fast to implement, and will require only a very small downtime window. <strong>This is what we’re going to discuss in this post.</strong></li></ol><p>To initialize a new Stolon cluster with existing PostgreSQL data, you’ll need to run through the following steps:</p><ol><li>Prepare the existing PostgreSQL database for Stolon.</li><li>Install an empty Stolon cluster (in init mode “new”) using helm (to get the generated postgresql.conf and pg_hba.conf files) and purge it after.</li><li>Install stolon (in init mode “existing”) with existing postgres folder, after replacing the above files with the ones generated by the first stolon installation.</li></ol><h3>1. Prepare the existing PostgreSQL database for Stolon</h3><p>If you currently have an existing postgres folder which stores all the data from your “old” PostgreSQL database, you’ll have to first “prepare” it to run under stolon, which basically means that there are 2 database roles needs to be created: a <strong>superuser </strong>and <strong>replication</strong>.</p><p>If you still have your old PostgreSQL database running, you can just create those roles and take the new “modified” postgres folder to stolon (move to step 2).</p><p>But if not, you can achieve this by spinning up a new PostgreSQL container (with the postgres folder mounted to it), create the required roles, and kill the container once done.</p><ul><li>First, modify your current postgres/pg_hba.conf file to enable all local connections (although not required, it’ll simplify things).</li><li>Start a new PostgreSQL container (in the same version as your current postgres version where you took the postgres folder from) and mount your postgres folder to it. Here is an example command to lunch such a container:</li></ul><pre># docker run -d -e PGDATA=/var/lib/pgsql/9.4/data/ -p 5432:5432 -v /&lt;PATH_TO_POSTGRES&gt;/postgres:/var/lib/pgsql/9.4/data --name &quot;postgres_for_stolon&quot; postgres:9.4 postgres</pre><ul><li>Attach to the container and connect to your database</li></ul><pre># docker exec -it postgres_for_stolon bash<br># psql -h localhost -Upostgres &lt;DB_NAME&gt;</pre><ul><li>Create the superuser and replication user roles (their password needs to match the password in Stolon’s helm chart, which we’ll talk about later)</li></ul><pre># create role &quot;stolonuser&quot; with login SUPERUSER encrypted password &#39;&lt;STOLON_PASSWORD&gt;&#39;;<br># create role &quot;repluser&quot; with login replication encrypted password &#39;&lt;REPL_PASSWORD&gt;&#39;;</pre><ul><li>Exit the container, and delete it (since no longer needed)</li></ul><pre># docker stop postgres_for_stolon<br># docker rm -f postgres_for_stolon</pre><ul><li>Copy the modified postgres folder to all other K8s nodes (this is needed since we’ll need the same postgres data directory on all nodes that will run the Stolon Keepers).</li></ul><h3>2. Install an empty Stolon cluster using helm</h3><p>We’ll now install an empty stolon cluster, to get the generated postgresql.conf and pg_hba.conf files, so we can use them later on.</p><ul><li>Clone stolon’s helm chart, which is documented here: <a href="https://github.com/helm/charts/tree/master/stable/stolon">https://github.com/helm/charts/tree/master/stable/stolon</a></li><li>Perform the necessary changes to the values.yaml file, and add your desired pgParameters and clusterSpec according to what is configured in your current PostgreSQL database.</li><li>Deploy the modified chart:</li></ul><pre># helm install &lt;PATH_TO_CHART&gt; --name stolon</pre><ul><li>Monitor the Keepers’ logs and wait for both the master and standby to be ready</li></ul><pre># kubectl logs -f stolon-keeper-0<br># kubectl logs -f stolon-keeper-1</pre><ul><li>Copy the generated postgresql.conf and pg_hba.conf files to a temporary location (on both K8s nodes which had the keepers running on)</li></ul><pre># cp /&lt;STOLON_DATA_DIR&gt;/postgres/postgresql.conf /&lt;STOLON_DATA_DIR&gt;/postgres/pg_hba.conf /tmp/</pre><ul><li>Delete stolon’s chart</li></ul><pre># helm del --purge stolon</pre><ul><li>Delete all contents inside Stolon’s directory (on both K8s nodes)</li></ul><pre># rm -rf /&lt;STOLON_DATA_DIR&gt;/*</pre><h3>3. Install stolon with existing postgres folder</h3><p>We’ll now copy our original postgres data directory to Stolon’s data directory, replace the postgresql.conf and pg_hba.conf files in the original directory with the ones generated before, and start stolon with init mode “<strong>existing</strong>”.</p><ul><li>Copy the existing postgres folder to Stolon’s folder (on both K8s nodes)</li></ul><pre># cp -r /&lt;PATH_TO_POSTGRES_FOLDER&gt;/postgres /&lt;STOLON_DATA_DIR&gt;/</pre><ul><li>Replace the modified postgresql.conf and pg_hba.conf files with the existing files in postgres folder (on both K8s nodes)</li></ul><pre># cp /tmp/postgresql.conf /tmp/pg_hba.conf /&lt;STOLON_DATA_DIR&gt;/postgres</pre><ul><li>Change permissions on Stolon’s folder (on both K8s nodes)</li></ul><pre># chown -R 1000:docker /&lt;STOLON_DATA_DIR&gt;/</pre><ul><li>Modify stolon’s create-cluster job yaml file to not create a new cluster database, but rather use an existing one (this theoretically can also be done by changing the values.yaml file of the chart or passing some arguments when deploying the chart, but I found it very tricky with this specific property)</li></ul><pre># sed -i -e &#39;s/&quot;initMode&quot;: &quot;new&quot;/&quot;initMode&quot;: &quot;existing&quot;, &quot;existingConfig&quot;: { &quot;keeperUID&quot;: &quot;keeper0&quot; }/g&#39;   &lt;STOLON_CHART_PATH&gt;/templates/hooks/create-cluster-job.yaml</pre><ul><li>Install Stolon with the modified chart</li></ul><pre># helm install &lt;STOLON_CHART_PATH&gt; --name stolon</pre><p><strong>A new Stolon cluster will be created (Keepers, Proxies, etc.) but the </strong><strong>postgres data dir will not be overwritten, and all your data will be available.</strong></p><p>You can monitor the initialization status the same as we did before, and once the master and standby keepers are ready, you can connect to postgres (locally from Stolon Keeper for example) and ensure that your data is still there.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=554072a2ba8e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/postgresql-high-availability-on-k8s-the-good-hard-way-554072a2ba8e">PostgreSQL High Availability on K8s — The good hard way</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ansible playbook — Creating elasticsearch cluster on ubuntu nodes using GCE discovery plugin]]></title>
            <link>https://medium.com/talking-tech-all-around/ansible-playbook-creating-elasticsearch-cluster-on-ubuntu-nodes-using-gce-discovery-plugin-20a7c03248a3?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/20a7c03248a3</guid>
            <category><![CDATA[google]]></category>
            <category><![CDATA[gce]]></category>
            <category><![CDATA[plugins]]></category>
            <category><![CDATA[ubuntu]]></category>
            <category><![CDATA[ansible]]></category>
            <dc:creator><![CDATA[Shahaf Yehezkel]]></dc:creator>
            <pubDate>Fri, 01 Feb 2019 10:49:17 GMT</pubDate>
            <atom:updated>2019-02-10T13:22:22.035Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/435/1*DGYeas1SUdV2JLgcl6Cqcg.png" /></figure><h3>Ansible playbook — Creating an elasticsearch cluster on ubuntu nodes using the GCE discovery plugin</h3><h4>Setting up an elasticsearch engine cluster is easy. I wanted to make it even easier and that is why I made this quick Ansible playbook.</h4><p>Before I start explaining about my playbook, let’s understand how Elastic and Google are helping us with the “Discovery Plugin”, well…</p><h4>What is the GCE discovery plugin?</h4><p>In order to create a cluster between a few nodes, the nodes need to “know” what other nodes are looking for a party, and we can manage it in a few ways. Today we are here to talk about one of the many ways, GCE Discovery Plugin.</p><p>Elastic has published a plugin for Google cloud servers. First, the plugin needs to be installed with the following command:</p><pre>sudo bin/elasticsearch-plugin install discovery-gce</pre><p>After that, we need to add the following lines to the elasticsearch.yml file (you can find it in /etc/elasticsearch if you did not make any changes during installation):</p><pre>cloud:<br>  gce:<br>      project_id: &lt;your-google-project-id&gt;<br>      zone: &lt;your-zone&gt;<br>discovery:<br>      zen.hosts_provider: gce</pre><ul><li>note that you can create a cluster with nodes even if they are in different zones, just use the following syntax:</li></ul><pre>zone: [&quot;us-east1-b&quot;, &quot;us-east1-c&quot;, &quot;us-west2-a&quot;]</pre><p>Now that we understand how GCE discovery plugin works, let’s see how I used it in my Ansible-playbook for creating a cluster within one simple command.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/416/1*w68vXPycPopfXh7np7_TZg.png" /><figcaption><strong>Ansible</strong> is the simplest way to automate apps and IT infrastructure</figcaption></figure><p>This playbook will perform three steps:</p><ol><li>Install elastic GPG-Key</li><li>Install Java</li><li>Install and config elasticsearch</li></ol><p>In this particular post, I chose to talk about step number three and how I used Ansible variables and of course the GCE discovery plugin settings for this playbook. You can take a look at my playbook in <a href="https://github.com/ShahafSW/Ansible_Elasticsearch_Cluster">Github</a>. There are a few things we need to notice when configuring a cluster:</p><h4>cluster.name:</h4><p>I used the Ansible group name from Ansible’s inventory as a variable.</p><h4>node.name:</h4><p>I used Ansible’s hostname as a variable.</p><h4>node.master:</h4><p>I used “true” (not as a variable), but this setting may be different depending on your cluster.</p><h4>node.data:</h4><p>I used “true” (not as a variable), but this setting may be different depending on your cluster.</p><h4>network.host:</h4><p>[_local_, _site_] (not as a variable)</p><h4>discovery.zen.minimum_master_nodes:</h4><p>“Elastic” recommends to use this equation:</p><pre>(master_eligible_nodes / 2) + 1</pre><p>After dividing the master_eligible_nodes by two, you need to floor the result, for example, 3/2=1.5 (floor the number to 1). I have three nodes in my cluster and they are all set to:</p><pre>node.master: true</pre><p>In this playbook, I am using the equation on all of the nodes as per “Elastic”’s recommendation.</p><h4>Now it is the part where the GCE discovery plugin needs to be set:</h4><h4>project_id:</h4><p>I am using Google’s API in order to get the nodes project id.</p><h4>zone:</h4><p>I am using Google’s API in order to get the node’s zone.</p><p>Please notice that you need to have the following Google’s permission to be able to query Google’s API for “project_id” and “zone”:</p><pre>--scopes=compute-rw</pre><p>*Note that granting these permissions will allow your server to make changes in your GCE.</p><h4>That’s it!</h4><p>We covered the important parts in my playbook which you can clone from the <a href="https://github.com/ShahafSW/Ansible_Elasticsearch_Cluster">git repository</a> and play it over your ubuntu nodes.</p><p>Don’t hesitate to ask questions if you have any, I will be glad to assist!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=20a7c03248a3" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/ansible-playbook-creating-elasticsearch-cluster-on-ubuntu-nodes-using-gce-discovery-plugin-20a7c03248a3">Ansible playbook — Creating elasticsearch cluster on ubuntu nodes using GCE discovery plugin</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to nail the CKA exam]]></title>
            <link>https://medium.com/talking-tech-all-around/how-to-nail-the-cka-exam-7944be5476a4?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/7944be5476a4</guid>
            <category><![CDATA[devops-as-a-passion]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[skywiz]]></category>
            <category><![CDATA[cka]]></category>
            <dc:creator><![CDATA[Lev@Skywiz.io]]></dc:creator>
            <pubDate>Sun, 21 Oct 2018 06:14:51 GMT</pubDate>
            <atom:updated>2018-11-01T05:23:11.953Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QGgiEu4rolD3JsOVt6UkAg.png" /></figure><p>Certified Kubernetes Administrator is, in my opinion, the most prestigious certification in the devops domain today. It’s like RHCE 10 years ago. The exam is hands-on, considered to be technically challenging, and it has tight time limit of 3 hours.</p><p>Ever since our senior solution architect, <a href="https://www.linkedin.com/in/itamary/">Itamar Yerushalmi</a>, passed his exam I have really wanted to get this certification for myself.</p><p>I passed successfully with score of 91% on the first try and want to share how I studied for it and what worked for me.</p><blockquote><em>Note: I passed this exam in Oct 2018. This is no longer an open book exam and only Kubernetes docs are allowed. Due to the NDA they make you sign, I cannot talk about questions.</em></blockquote><p>I split this article into two sections<strong> learning sources</strong> and <strong>practice advices</strong>.</p><h3>Learning sources:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*kSj1JWxOFo8fKMEM.jpg" /><figcaption>Yeah…That….</figcaption></figure><ul><li><a href="https://training.linuxfoundation.org/training/kubernetes-fundamentals/">Linux Foundation</a> — In my opinion, this is by far the best source to learn from. As the official training for the CKA exam, it covers almost everything that was needed to pass the exam. Toward the end of the course it becomes less in-depth in subjects like helm or federation. It’s a bit pricey regularly, but we purchased this training for $150 during the Black Friday discount. One thing that I dislike about this source is the interface — it feels like it was done in 1997 and they didn’t bother to change it ever since.</li><li><a href="https://linuxacademy.com/linux/training/course/name/kubernetes-the-hard-way">Linux academy ‘kubernetes the hard way’ </a>— Linux academy is a great source for many courses. I used it for my AWS Solution architect exam, I Google Cloud Architect Exam and I suggest it to everyone. This course is even better that the original Kubernetes the hard way document. You get even more detailed explanations, labs and even great smoke tests to check that your cluster is actually working.</li><li><a href="https://cloudacademy.com/learning-paths/certified-kubernetes-administrator-exam-preparation-242/">Cloud academy</a> — In general, this one is okay. They have good material, but not enough and a bit outdated. They do have practice labs with different cloud environments — Azure, AWS — which is nice.</li><li><a href="https://linuxacademy.com/linux/training/course/name/certified-kubernetes-administrator-preparation-course">Linux Academy CKA</a> — <strong>Avoid learning from this course</strong>. The course covers irrelevant subjects, such as installation of Centos k8s cluster with kubeadm; goes into unnecessary details, like 12 minutes video covering all available persistent volume types; and doesn’t go in-depth enough on some important subjects like security, for one example. Quizzes include questions that are not even covered in the videos. This course proved to be a very frustrating experience.</li><li><a href="https://www.youtube.com/watch?v=4XAlFEVKGOw&amp;list=PL5cnXKhXmBeZlQSe6xBSchFCM2Mi3fYCa">CKA Office Hours</a> — This is a YouTube video series. Some of the videos are a bit long and boring but it does give examples of real questions and what the environment looks like.</li><li><a href="https://www.youtube.com/watch?v=0Omvgd7Hg1I">Life of a packet video</a> — Also on YouTube, it is very important to understand how networking in K8s works even if you’re not aiming for certification it explains a bit about K8s magic.</li></ul><h3>Practice advices:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/346/0*BPyMNvJGR0PJUsLW.jpg" /></figure><ul><li>Do everything manually — Don’t copy&amp;paste anything. Type all the Yaml files, type all the commands, type all the file configurations. <strong>Typing leads to typos, typos lead to troubleshooting, troubleshooting leads to better understanding.</strong></li><li>Make sure you do Kubernetes the hard way. And when I say do it I mean actually do it by typing everything manually. It’s one thing to copy paste cfssl json and completely another thing to type it, make a mistake, figure it out 2 hours later from the kubelet logs that can’t register to the Api server. It makes you feel like a real K8s ninja.</li><li><a href="https://www.katacoda.com/joshuatalb/courses/kubernetes">Katacoda</a>— While most of the scenarios are crap, the troubleshooting scenarios are great!!! Especially if you don’t have any real K8s experience.</li><li><a href="https://play.instruqt.com/public">Instruqt</a>.com — has some nice Kubernetes scenarios.</li><li><a href="https://linuxacademy.com/linux/training/course/name/certified-kubernetes-administrator-preparation-course">Linux academy CKA</a> — While it’s a horrible source to learn from as I mentioned above — it’s really great source of practice labs. And it’s even has a practice exam which is similar “spirit” to the real thing and I strongly suggest taking it.</li></ul><h3><strong>General Tips:</strong></h3><p>I would say that the exam is structured like that:</p><p><em>30% easy and obvious questions.</em></p><p><em>30% a bit more challenging question but nothing you can’t solve with enough time.</em></p><p><em>30% questions that require you to read and learn something unexpected fast.</em></p><p><em>10% questions that make sure that only a small amount of people can score100%.</em></p><p>So tips:</p><ul><li>You have only 3 hours — you must manage it well — questions that feels like they will take more than 10 minutes — write them down in the notepad ( <strong>in the exam tab notepad </strong>), and move on. Get back to them when all the rest is done. I had 4 questions and I went back to them after I finished all the rest within 1.5 hours.</li><li>Make sure you know how to type autocompletion command by heart. <strong><em>source &lt;(kubectl completion bash)</em></strong><em>.</em><strong><em> </em></strong>Everything is easier with double tab.</li><li>Make sure you know what error looks like when your yaml is incorrect and how to use <strong><em>kubectl explain</em></strong>. Going to k8s documentation is time consuming and time is of an essence.</li><li>There is a great Heptio pro tip that gives an example of how to bootsrap a yaml file fast. You should know it by heart it will save you a lot of time. You are not allowed to copy&amp;paste more than 1–2 lines. For example:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2wRWsQ3WOwEeH88mAvfSWg.png" /></figure><ul><li>Once you have yaml file created you can copy it, change it, reapply it. Don’t forget that you can use <strong>kubectl edit </strong>— to edit resources faster — therefore saving time on running kubectl apply commands :)</li><li>Check K8s what version will be in the exam (you can find in the candidate book) and make sure you learn with proper material. Features move from alpha to beta to v1 or become deprecated.</li></ul><p>Some of the things that I mention are covered in other articles. The point here was to summarize what I found was the most helpful for me during the exam and I hope you found it helpful and saves you a bit of time.</p><p>Good luck with your exam ! You can do it!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/435/0*98CnIagT_hTGYKHW.gif" /><figcaption>CKA is walking</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7944be5476a4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/how-to-nail-the-cka-exam-7944be5476a4">How to nail the CKA exam</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[HashiDays 2018 summary]]></title>
            <link>https://medium.com/talking-tech-all-around/hashidays-2018-summary-44ba08443c92?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/44ba08443c92</guid>
            <category><![CDATA[skywiz]]></category>
            <category><![CDATA[hashicorp]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[hashidays]]></category>
            <dc:creator><![CDATA[Lev@Skywiz.io]]></dc:creator>
            <pubDate>Sun, 01 Jul 2018 05:22:58 GMT</pubDate>
            <atom:updated>2021-05-14T06:45:57.886Z</atom:updated>
            <content:encoded><![CDATA[<p>TLDR If you’re considering going to HashiDays you should definitely go, it’s worth it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tC0i3ky0Hh_ixF4Lx7fcCw.jpeg" /><figcaption>400+ people — friendly and cozy</figcaption></figure><p>So let me set the scene for you. In 2014, I left the IT industry. At the time I was an IT manager responsible for hundreds of bare-metal servers, in multiple data-centers across several continents. I was pretty good with scripting, automation, and configuration managers. I had 10 years of experience as linux sysadmin for high traffic web platforms (Amobee, Taboola, FlashNetworks).</p><p>I left all that to start a sharing economy startup with my friends. My main responsibility in the startup was frontend and mobile (hybrid) development.</p><p>After successfully failing at that venture I decided to return to the IT world as a part time freelancer for an Israeli unicorn company called <a href="https://www.appsflyer.com/">Appsflyer</a>. I can’t even describe how lucky I was that Appsflyer team took me on-board the rocket-ship. I still remember saying “Devops is just a linux admin who can write some code” — thank God, <a href="https://www.linkedin.com/in/ariel-moskovich-2530aa18/">Ariel</a>, the Appsflyer Devop Lead and a brilliant guy in general, ignored my non-sense.</p><p>So there I was — back into the IT world after missing 2 years and saying “What the hell happened here?” Automatic builds and deployments, secret management, spot instances and auto-scaling in AWS, service discovery and configuration management with key/value store.</p><p>That’s where I learned about Vault&amp;Consul. For me it felt like Hashicorp were aliens from the future and, to be honest, I still think that. They just knew how IT was supposed to be designed and work even before IT professionals understood it. It felt right, it made sense, it was secured and cloud-ready, and it didn’t feel limiting.</p><p>So for me meeting developers of those amazing tools — Michell Hashimoto &amp; Hashi-team — is a dream come true.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*28d3pA_ztv8EoQb2dLbQ-g.jpeg" /><figcaption>I’m such a fanboy</figcaption></figure><h4><strong>One Day Consul Training</strong></h4><p>The training felt like it was meant for beginners to medium experience level. It started with basics like setups, service discovery and of course key/value storage. Then it moved into raft and gossip protocols, Serf and some nice features like cloud-integration and tag-based bootstrapping. We covered tooling like consul-template, a bit of fabio and envconsul. The main thing for me during the training was the fact that we’re close to the actual developers so we could get as deep as we want and get answers at the design and code level. By the way, the training is hands-on so you get to practice with a real server and with a real consul cluster. I just wish there was a separate training for Consul advanced topics and maybe some troubleshooting and disaster recoveries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IBzl_w5igH41pmPkJE_-0Q.jpeg" /><figcaption>This is so awesome — it’s an arcade machine to test your hashi-skills by instruqt</figcaption></figure><h4><strong>Overall, the Conference was great!</strong></h4><p>Here’s why, in my opinion, Hashicorp conferences are worth attending:</p><ol><li>Be the first to know — announcements like <a href="https://www.consul.io/intro/getting-started/connect.html">Consul connect</a> with the whole ideology, philosophy and design decisions. You get to hear it first and learn about it from the source. Roadmaps ( without timeline commitments — cause that’s how they roll ). Also, you free up your time and attention to learn those new features and products.</li><li>Best Practices — Hashicorp is not a consulting company so they tend to steer clear of suggesting best practices. So the way to learn about best practices is actually hearing about customers implementing large scale, beautiful and sophisticated solutions. More than 50% of the conference was by non-Hashicorp employees sharing their knowledge and experience.</li><li>Get inspired — talking to Armon or Mitchell, Alex or Clint Shryock is exciting. They make you want to try new things and improve your current solutions and drive-test alpha features. Share your experience with the tools with others members of the community.</li><li>Connect — You work with tools written by other humans, maybe some aliens too, and actually used by humans. You connect directly to Hashicorp developers, contributors, customers, system integrators, and other people who you will need the most when things are not working as they should. Actually meet contributors that are willing to mentor and on-board you to the project.</li><li>No aggressive marketing — I would say 97% of the conference was talking about open source features and products. The other 3% that refer to their enterprise mentioned for specific use cases where it’s relevant and even this was in “semi-apologetic” way.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/891/1*FQM3NnbcjCROwWpqhsHnag.jpeg" /><figcaption>They even made a consul macaroon!!!!</figcaption></figure><p>So there you go: Attend it! If you can’t fly get the streaming and watch it, and connect with people via Slack.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=44ba08443c92" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/hashidays-2018-summary-44ba08443c92">HashiDays 2018 summary</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Make Interviews Great Again!]]></title>
            <link>https://medium.com/talking-tech-all-around/make-interviews-great-again-c46037d1b7e2?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/c46037d1b7e2</guid>
            <category><![CDATA[sysadmin]]></category>
            <category><![CDATA[technical-interview]]></category>
            <category><![CDATA[linux]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[interview]]></category>
            <dc:creator><![CDATA[Lev@Skywiz.io]]></dc:creator>
            <pubDate>Tue, 24 Apr 2018 11:15:06 GMT</pubDate>
            <atom:updated>2018-04-25T11:35:01.114Z</atom:updated>
            <content:encoded><![CDATA[<p>Hiring is a challenge. Especially if you are a small company, ‘cause every single person counts. Even more so when trying to find a good linux SysAdmin.</p><p>While in general it’s a great idea to get to know a candidate before hiring, there’s very little you can do to ACTUALLY get to know someone. It takes time and several urgent 2am maintenances outages.</p><p>In my opinion, the primary goal of a technical interview is to determine whether your candidate can learn and whether he will “dig” your company’s culture and values.</p><p>Let me throw out a cliché line and tell you something that you’ve already heard — technical interviews are broken.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*fuIoRt9wNisQFEzgM0siZQ.png" /><figcaption>Linux SysAdmin interviews in 2018</figcaption></figure><p>Based on my experience, the most popular tools for technical interviews are question-based interviews, tests, exams, and “homework.” Let’s talk about each of these.</p><h4>Question-based interviews</h4><p>Most experienced candidates (that is, experienced in interviews, but not necessarily in technical knowledge) are familiar with questions like these:</p><ul><li>What is load average ? ( one of the best answers btw is <a href="http://www.brendangregg.com/blog/2017-08-08/linux-load-averages.html">here</a> )</li><li>Tell me how you would troubleshoot performance issues?</li><li>Describe a complex problem and how you solved it. Tell me about a project you’ve worked on that you are proud of. (Many candidates have pre-baked answers for that one.)</li></ul><p>Those questions are great, but not for a technical interview. And don’t get me started on specific scenarios that interviewers have encountered during their careers and are now asking as interview questions:</p><ul><li>How would you set up this specific nodejs code in a docker swarm orchestrator?</li><li>What potential issues might there be with web-sockets?</li><li>How would you solve my issues?</li></ul><p>It probably took hours or days of troubleshooting, trial and error, and elimination to find and fix that issue — the candidate may have had the same issue, or may have no idea how to solve the issue without logs, code, and setup under the pressure of the interview.</p><p><strong>“The homework”</strong></p><p>A few years ago it used to be part of the interview that took 1–2 hours. You got to install something like munin server and agent. Maybe even setup some dashboard, or write an access log parser and write to database.</p><p>Then, I stared to hear that candidates were assigned to set up a highly available mongodb cluster with a configuration management tool on Google or AWS.</p><p>Ok, so, first of all, companies should actually pay for the implementation of mongo cluster setup automation, even if there are plenty of playbooks, recipes, and other veggies to do it in 1 hour.</p><p>And what kind of a test is that anyway?</p><p>Are you testing whether a candidate can find the correct digital ocean page in Google that describes this process, or that the candidate can find the right ansible playbook?</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FoYZEYsOeRDM%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DoYZEYsOeRDM&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FoYZEYsOeRDM%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/d5e4b249b520d40324436c6902ae702c/href">https://medium.com/media/d5e4b249b520d40324436c6902ae702c/href</a></iframe><p>Tests and Exams</p><p>Personally, I think that tests are great to pre-filter candidates. This saves time for both you and your potential hires.</p><p>I like when someone is given a task that is not copy-pastable from stackoverflow — something that proves that the candidate can search and find the right information, learn some basics quickly, and then implement that information as required by the task.</p><p>We wanted to create that experience on steroids, and if possible, to remove interview stress, add some fun to the interview, and add a touch of our company culture.</p><p><strong>Introducing Escape Box</strong></p><p>Wikipedia says, “An <strong>escape room</strong> is a physical adventure game in which players solve a series of puzzles and riddles using clues, hints and strategy to complete the objectives at hand.”</p><p>We took that concept and replaced the room with a server, and in our box, the riddles and puzzles became a bit more technical. And there you go — we have an Escape Box.</p><p>First, you will receive an email with an IP and GPG key. If you browse to that IP, you will see something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pcmHjH0KlNa0YFc0Yn3VvA.png" /></figure><p>Hopefully you will try to SSH to the server and this will lead to this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/976/1*DZ_4hZspcKn-mv6BNDYeOw.png" /></figure><p>But examining a bit closer you might find what is really running on port 22:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*1YMGYzTQGWF8UvjaSf_lPg.png" /></figure><p>So, after struggling a bit, completing another step, and finally logging in to the system, you will be greeted by this message:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*idFAohUzfoUV7YQBRMBP0A.png" /></figure><p>Once elasticsearch is installed a background process identifies that port 9200 is now open and throws a broadcast message detailing the next step… and so it goes until the last — point encrypted file. Which is decrypted by the GPG key the candidate initially got.</p><p>Here’s what our latest and greatest candidate thinks:</p><blockquote>Great virtual escape room game, good challenge using linux real world tools.</blockquote><blockquote>The riddles were challenging but fair, I would polish the instructions a bit.</blockquote><blockquote>I liked it and I would take part in similar games.</blockquote><blockquote>Fun experience, highly recommended.</blockquote><p>Sure — it’s far from perfect, but here’s where it gets even better. If you successfully complete the Escape Box and are hired, you must do one of the following:</p><ol><li>Improve something in the previous box</li><li>Add a new step, puzzle, or riddle to the box</li></ol><h4>So what’s next?</h4><p>After discussing our idea with several friends, we decided to share it with the world. We’ve decided to migrate this to automated vagrant installation, and will be sharing the Escape Box in a public repository so that everyone can create new boxes and contribute recipes, puzzles, and riddles.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c46037d1b7e2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/make-interviews-great-again-c46037d1b7e2">Make Interviews Great Again!</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kafka Multi-Tenancy Architecture: SSL client authentication]]></title>
            <link>https://medium.com/talking-tech-all-around/how-to-enable-and-verify-client-authentication-in-kafka-21e936e670e8?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/21e936e670e8</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[authentication]]></category>
            <category><![CDATA[pki]]></category>
            <category><![CDATA[kafka]]></category>
            <category><![CDATA[ssl-certificate]]></category>
            <dc:creator><![CDATA[Itamar Yerushalmi]]></dc:creator>
            <pubDate>Tue, 20 Feb 2018 11:14:23 GMT</pubDate>
            <atom:updated>2018-03-01T14:32:51.712Z</atom:updated>
            <content:encoded><![CDATA[<p>Kafka is a distributed streaming platform. As such, we would want many consumers and producers to write to the same Kafka cluster.</p><p>In one of our latest projects in a large scale cyber-security company, we created a large scale deployment of a Kafka cluster that needed to serve approximately 10TB of digested data each day, split across over 100 million events. This was done in order to limit each customer’s ability to access only their topic in a multi-tenancy architecture.</p><blockquote>There are a few guides over the internet to enable SSL client authentication on Kafka. Some of them are outdated or incomplete, while the others include overkill Kerberos implementation. So I created this one, which is based on Kafka’s official documentation at this <a href="http://kafka.apache.org/documentation.html#security_ssl">link</a>, hopefully you’ll find it useful.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bkUDyya838QHOrxY7cGglg.jpeg" /><figcaption>Producer can write to all topics. Consumer can access secured topic only with right certificate.</figcaption></figure><p><strong>For this example, I have one Kafka server and 2 clients (consumers).</strong></p><p>Server Side:</p><p>Make sure that you have truststore and keystore JKSs for each server.</p><ol><li>In case you want a self signed certificate, you can use the following commands:</li></ol><pre>export PASSWORD=password</pre><pre>keytool -keystore kafka.server.keystore.jks -alias localhost -validity 365 -genkey</pre><pre>openssl req -new -x509 -keyout ca-key -out ca-cert -days 365</pre><pre>keytool -keystore kafka.server.truststore.jks -alias CARoot -import -file ca-cert</pre><pre>keytool -keystore kafka.client1.truststore.jks -alias CARoot -import -file ca-cert</pre><pre>keytool -keystore kafka.client2.truststore.jks -alias CARoot -import -file ca-cert</pre><pre>keytool -keystore kafka.server.keystore.jks -alias localhost -certreq -file cert-file</pre><pre>openssl x509 -req -CA ca-cert -CAkey ca-key -in cert-file -out cert-signed -days 365 -CAcreateserial -passin pass:$PASSWORD</pre><pre>keytool -keystore kafka.server.keystore.jks -alias CARoot -import -file ca-cert</pre><pre>keytool -keystore kafka.server.keystore.jks -alias localhost -import -file cert-signed</pre><pre>keytool -keystore kafka.client1.keystore.jks -alias localhost -validity 365 -genkey</pre><blockquote>Make sure that you configure the certificate properly — this will be used later in the ACL configuration.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ccmX5gp_Hs2Q9oQnDduDzQ.png" /></figure><pre>keytool -keystore kafka.client1.keystore.jks -alias localhost -certreq -file cert-file</pre><pre>openssl x509 -req -CA ca-cert -CAkey ca-key -in cert-file -out cert-signed -days 365 -CAcreateserial -passin pass:$PASSWORD</pre><pre>keytool -keystore kafka.client1.keystore.jks -alias CARoot -import -file ca-cert</pre><pre>keytool -keystore kafka.client1.keystore.jks -alias localhost -import -file cert-signed</pre><pre>keytool -keystore kafka.client2.keystore.jks -alias localhost -validity 365 -genkey</pre><pre>keytool -keystore kafka.client2.keystore.jks -alias localhost -certreq -file cert-file</pre><pre>openssl x509 -req -CA ca-cert -CAkey ca-key -in cert-file -out cert-signed -days 365 -CAcreateserial -passin pass:$PASSWORD</pre><pre>keytool -keystore kafka.client2.keystore.jks -alias CARoot -import -file ca-cert</pre><pre>keytool -keystore kafka.client2.keystore.jks -alias localhost -import -file cert-signed</pre><p>2. On Kafka servers — <strong>server.properties</strong> — add the following lines:</p><pre>ssl.enabled.protocols=TLSv1.2,TLSv1.1,TLSv1</pre><pre>ssl.endpoint.identification.algorithm=HTTPS</pre><pre>ssl.keymanager.algorithm=SunX509</pre><pre>ssl.keystore.location=/PATH-TO/kafka.server.keystore.jks</pre><pre>ssl.keystore.password=password</pre><pre>ssl.keystore.type=JKS</pre><pre>ssl.protocol=TLS</pre><pre>ssl.trustmanager.algorithm=PKIX</pre><pre>ssl.truststore.location=/PATH-TO/kafka.server.truststore.jks</pre><pre>ssl.truststore.password=password</pre><pre>ssl.truststore.type=JKS</pre><pre>authorizer.class.name=kafka.security.auth.SimpleAclAuthorizer</pre><pre>allow.everyone.if.no.acl.found=true</pre><p>3. Add the following to the <strong>advertised.listeners</strong> and the listeners keys SSL values on port 9092:</p><pre>advertised.listeners=PLAINTEXT://KAFKA_IP:9092,SSL://KAFKA_IP:9093</pre><pre>listeners=PLAINTEXT://KAFKA_IP:9092,SSL://KAFKA_IP:9093</pre><p>4. Change</p><pre>ssl.client.auth=none</pre><p>To</p><pre>ssl.client.auth=required</pre><p>5. For logging — enable DEBUG on Kafka authentication by changing the line in <strong>log4j.properties —</strong></p><p>From:</p><pre>log4j.logger.kafka.authorizer.logger=WARN, authorizerAppender</pre><p>To:</p><pre>log4j.logger.kafka.authorizer.logger=DEBUG, authorizerAppender</pre><p>6. On the client servers add the following lines to<strong> consumer.properties</strong> and <strong>producer.properties</strong>:</p><pre>security.protocol=SSL</pre><pre>ssl.truststore.location=/PATH-TO/ssl/kafka.client1.truststore.jks</pre><pre>ssl.truststore.password=password</pre><pre>ssl.keystore.location=/PATH-TO/kafka.client1.keystore.jks</pre><pre>ssl.keystore.password=password</pre><pre>ssl.key.password=password</pre><p>Notice that you are using different files for the different servers.</p><p>7. Produce messages to a new topic:</p><pre>./bin/kafka-console-producer.sh — broker-list KAFKA_IP:9093 — topic secured-topic — producer.config /opt/kafka/config/producer.properties</pre><pre>&gt;1</pre><pre>CTRL+D</pre><p>You should now be able to write to the topic from both servers — if you look at the <strong>kafka-authorizer.log</strong> on the server you will see the following messages:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O3Vi8Q2dJZMp0Bn41Ty9kA.png" /></figure><p>8. Using the kafka-acl command — set permissions for one of the users — by default, it will remove permissions from the other user:</p><pre>bin/kafka-acls.sh — authorizer kafka.security.auth.SimpleAclAuthorizer — authorizer-properties zookeeper.connect=ZOOKEEPER_IP:2181 — add — allow-principal User:”CN=client1,OU=1,O=1,L=1,ST=1,C=1&quot; — operation Write — operation Read — operation Describe — topic secured-topic</pre><p>9. Now, if you try to write or read from the granted user, you will succeed —</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-z0z75zhge5dfRHcefPaog.png" /></figure><p>and from the other user — permission denied.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7hd8E50pU5myRF1toK4tow.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=21e936e670e8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/how-to-enable-and-verify-client-authentication-in-kafka-21e936e670e8">Kafka Multi-Tenancy Architecture: SSL client authentication</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ELK6.x on Centos7.x — Quick install]]></title>
            <link>https://medium.com/talking-tech-all-around/elk6-x-on-centos7-x-quick-install-219dd0a97065?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/219dd0a97065</guid>
            <category><![CDATA[elasticsearch]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[elk]]></category>
            <category><![CDATA[google-cloud-platform]]></category>
            <category><![CDATA[automation]]></category>
            <dc:creator><![CDATA[Itamar Yerushalmi]]></dc:creator>
            <pubDate>Mon, 29 Jan 2018 09:44:32 GMT</pubDate>
            <atom:updated>2018-01-29T10:57:32.669Z</atom:updated>
            <content:encoded><![CDATA[<p>How many times did you want to to install something for testing something specific. “Let’s do it quick manually , i won’t need it more then this time” is something i always negotiate with myself . As part of working a-lot with ELK i come across with this question a-lot. This time i was looking for a fast way to install one node -ELK6 for Centos/RHEL 7.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*vEb40HRegQgqJvz7Qz8zZg.png" /></figure><p>Instead of doing the work myself i started to look for something on the web but there are a lot of out dated playbooks online and github has become a graveyard for not-operational stuff .I have based my playbook on<a href="https://github.com/bingoarun/ansible-elk"> Arun’s Prasath</a> one but did some minor changes to fit the ELK 6.x . The playbook is written as a role — so you can use it in your <strong>Ansible CM.</strong></p><p><strong>You can find it here:</strong></p><p><a href="https://github.com/skywiz-io/Ansible-ELK">skywiz-io/Ansible-ELK</a></p><p>I’ll be posting a second part on “Cluster mode” as well in the next weeks.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=219dd0a97065" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/elk6-x-on-centos7-x-quick-install-219dd0a97065">ELK6.x on Centos7.x — Quick install</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[VMworld17 It was …. Nice.]]></title>
            <link>https://medium.com/talking-tech-all-around/vmworld17-it-was-nice-4c78aa22a0ad?source=rss----2dd325eb3f7a---4</link>
            <guid isPermaLink="false">https://medium.com/p/4c78aa22a0ad</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[cloud]]></category>
            <category><![CDATA[vmware]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <category><![CDATA[aws]]></category>
            <dc:creator><![CDATA[Tal Shatz]]></dc:creator>
            <pubDate>Tue, 26 Sep 2017 14:18:50 GMT</pubDate>
            <atom:updated>2017-09-26T14:18:50.134Z</atom:updated>
            <content:encoded><![CDATA[<p>As I wrote in my last post (<a href="https://medium.com/talking-tech-all-around/vmworld17-europe-on-the-road-49283c7bcd1d"><strong>Check it out</strong></a><strong> ) </strong>this was my first time in #VMworld so I didn’t know what to expect to from the venue itself, sessions, the audience, etc. My mission was to meet interesting people and hear interesting stuff in the sessions.Let’s say that the people were definitely interesting, rather the sessions slightly less.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O6CLt85bPacI6fEgQH-Euw.jpeg" /></figure><p>If we start with the sessions — I did expect to get more out and see how we can leverage the new announcements and how our customers can benefit from these announcements.</p><p>VMware announced, as you already know, about the cooperation with #AWS,</p><p><strong>What does it actually mean?</strong> The offering of AWS &amp; VMware will be pushed towards customers by the VMware team — that means that AWS team is irrelevant here from the business side. The solution means that you will get a full SDDC solution (vCenter, vSphere, vSAN, NSX) that holds four (4) physical nodes as a minimum and can scale up to sixteen (16) , you can have a full SDDC after two (2) hours only. However, this solution is currently available only in <strong>Oregon, USA</strong></p><p>They are already talking about expanding it and by the beginning of 2018 it will be available in other regions in the US in late 2018 it will be available in Europe as well.</p><p><strong>What do we need it for? Using this solution will enable you</strong> to scale-out to the public cloud from your private cloud via VMware vRealize Automation, it will make it a real #hybrid #cloud solution. In the next release, you will be able also to create L2 VPN between onprem and public cloud which will you to vMotion the machines to the cloud and back.</p><p>Not like VMware, you can’t oversubscribe in the public cloud, so if one of your AWS Instances isn’t utilizing the entire stack of this resource it’s basically money thrown out.In addition, usually with on-prem you are utilizing 50–60 % of the resources in order to keep your growth factor and in case you will need failover your VMs, while using SDDC in AWS you can scale-out your SDDC by spinning on a vSphere host and automatically it will kick-in under 10 minutes. Therefore, you can utilize more of the cluster most of the time and keep it on a daily basis around 90%. Furthermore, think about having an HA automatically adding a new host to the cluster while keeping the old one decommissioned if it’s HW failure.</p><p><strong>What about using AWS Services with Vmware</strong>? All of #AWS services are available for usage through the VMs on the VMware side. Think about having multiple VMware VMs like databases that are legacy types and can leverage the EC2 Spot instances or even ECS.</p><p>That’s impressive. And finally, but you didn’t hear it from me …. I’ve heard that possible EBS (Amazon Elastic Block Store) will be the storage for the VMware vSphere cluster, the NvME I’m guessing will be the caching layer in the near future J. An additional feature, Again …NOT heard it from me, vRealize Automation will be able to manage also the VMware On-prem, VMware SDDC on AWS &amp; AWS web services side, you will get a single dashboard with the same look and feel.</p><p>The announcements on <strong>#Pivotal #PKS</strong> on top of VMware was slightly disappointing, I mean they are investing lots in the Cloud Foundry where it should be an enterprise platform and now in addition also they are going to put their money on Kubernetes project which is a most used container orchestration I believe and becomes a type of standard, but it looked like that PKS &amp; PCF are overlapping which brings me to why continue the PCF for long term . From the other hand, the Integrated containers future looks good and to be able to run VIC along side a VM will be very good for IT staff management. I see a way where the PKS and #VIC will somehow work together in the future.</p><p>I do believe that VMware has a long vision of how they will more and more leverage this, so let’s wait and see.</p><p><strong>VMware Integrated OpenStack (VIO)</strong> had also new announcements, and I do feel they are interesting. They have added Containers support users that can run VMs alongside containers on VIO.. By using VMware vRealize Integration with VIO, you can now embed OpenStack components in blueprints. OpenStack deployments can now be managed via the Horizon UI as an independent tab in vRA. You can scale under a single VIO deployment, multiple vCenter for large-scale environments. In addition, they have added support for live VM resize (CPU , Memory &amp; disk) without the need to shut down the VM , they also added Firewall as a service and much more</p><p>There is something interesting I want to share with you. I’ve gone through the booths to see what’s new and if there are any special new vendors in the show. I found out something amazing that I can’t understand why. I saw that lots of well-known storage companies attended and I try to figure out why? I’m talking about companies such as Infinidat, NetApp, PureStorage , HDS, Kaminario, Qnap and of course EMC ,HP all these companies took part in the show, why I was wondering? Because I couldn’t find what is the new message these companies are bringing to the table. It wasn’t what I expected from them but as I said before “I didn’t come with expectations” J</p><p>In a quick look, #Rubrik made a great scene around their product and they have put up a nice show, I don’t know if the product is quite there yet but they are here to make noise for sure. From the other hand, #Veeam which is usually my favorite vendor didn’t bring to many major announcements that you can say “WOW”. The CDP Feature is a good thing and it’s about time they did it and of course the NAS backup which is a must today to compete in the Enterprise products with the giants. Other than that I wasn’t blown away, but let us see the final release and see how they will surprise us.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*_8JrNSE6f-GLXF72.png" /></figure><p><strong>For conclusion,</strong> I do see a bright future for VMware with the AWS collaboration, there is much potential here and I believe it will actually balance the market more from companies that went full-in the cloud and companies that where on-prem only. This is my highlight of course from VMworld this year.</p><p>See you next year <strong>#VMworld18</strong>, it was a great experience meeting amazing people, meet some clients and frankly also enjoy the wonderful city of Barcelona.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4c78aa22a0ad" width="1" height="1" alt=""><hr><p><a href="https://medium.com/talking-tech-all-around/vmworld17-it-was-nice-4c78aa22a0ad">VMworld17 It was …. Nice.</a> was originally published in <a href="https://medium.com/talking-tech-all-around">Talking Tech Around</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>