<?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[Stories by Satya Kumar on Medium]]></title>
        <description><![CDATA[Stories by Satya Kumar on Medium]]></description>
        <link>https://medium.com/@satya.kumar?source=rss-749081e90d99------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*wscEeJ5RvpZao12P9Q8RPg.jpeg</url>
            <title>Stories by Satya Kumar on Medium</title>
            <link>https://medium.com/@satya.kumar?source=rss-749081e90d99------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 22:03:07 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@satya.kumar/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[ Debugging “poetry: command not found” in Dockerized CI/CD Pipelines]]></title>
            <link>https://medium.com/@satya.kumar/debugging-poetry-command-not-found-in-dockerized-ci-cd-pipelines-6a4f436b0c26?source=rss-749081e90d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/6a4f436b0c26</guid>
            <category><![CDATA[sre]]></category>
            <category><![CDATA[dockerfiles]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[makefile]]></category>
            <dc:creator><![CDATA[Satya Kumar]]></dc:creator>
            <pubDate>Wed, 29 Oct 2025 02:37:38 GMT</pubDate>
            <atom:updated>2025-10-29T03:02:33.857Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’ve ever built Python projects inside Docker during a CI/CD pipeline, chances are you’ve hit a frustrating error like this:</p><pre>bash: poetry: command not found</pre><p>Even though… when you docker exec into the container and run the exact same command manually — it works perfectly. 🤯</p><p>This happened to me recently in a Jenkins pipeline, and I want to document the <strong>why</strong>, the <strong>debugging process</strong>, and the <strong>simple fix</strong> that finally solved it.</p><h3>🧪 The Setup</h3><p>I had a CI pipeline that:</p><ul><li>Spins up a container from a pre-baked Python base image</li><li>Uses Poetry to build a Python wheel</li><li>Publishes it to JFrog Artifactory with Twine.</li></ul><p>The image already had poetry and twine installed.</p><p><strong>Locally inside the container</strong>:</p><pre>docker exec -it cdp-auth-lib-builder bash<br>root@container:/app# poetry build --format wheel<br>Building ...<br>✅ Wheel built successfully</pre><p><strong>But in the Jenkins pipeline</strong>, this failed:</p><pre>docker exec -i cdp-auth-lib-builder bash -lc &#39;poetry build --format wheel&#39;<br>/usr/bin/bash: line 1: poetry: command not found</pre><p>Same command, different behavior. Why?</p><h3>🧠 The Real Reason</h3><p>When you exec into a container interactively (docker exec -it), it launches a <strong>login + interactive shell</strong>, which sources:</p><ul><li>/etc/profile</li><li>~/.bashrc (or ~/.profile)</li></ul><p>Those files typically <strong>set up the PATH</strong> to include places like:</p><pre>/opt/venv/bin<br>/root/.local/bin<br>/usr/local/bin</pre><p>That’s where Poetry and Twine live in most Python images.</p><p>But when Jenkins runs:</p><pre>docker exec -i &lt;container&gt; bash -lc &quot;poetry build --format wheel&quot;</pre><p>…it launches a <strong>non-interactive shell</strong>.<br> No ~/.bashrc. No PATH adjustments. No poetry.</p><p>👉 So the binaries are still in the container, but the shell doesn’t know where to find them.</p><h3>🔍 Quick Debug Tip</h3><p>You can verify this with:</p><pre>docker exec -it &lt;container&gt; bash -lc &#39;echo $PATH&#39;</pre><p>and compare it to:</p><pre>docker exec -it &lt;container&gt; bash<br>root@container:/app# echo $PATH</pre><p>You’ll see they’re different.</p><h3>🪄 The Simple Fix</h3><p>All we needed to do was <strong>explicitly prepend the directories</strong> where Poetry and Twine are installed, before running the commands in our Makefile:</p><pre># Typical Poetry/Twine locations in Python images<br>EXEC_PATHS := /opt/venv/bin:/root/.local/bin:/usr/local/bin</pre><p>Then in the docker exec calls:</p><pre>docker_build_whl:<br>	docker exec -i $(CONTAINERNAME) bash -lc &#39;\<br>	  export PATH=&quot;$(EXEC_PATHS):$$PATH&quot;; \<br>	  cd &quot;$(APP_SUBDIR)&quot;; \<br>	  poetry build --format wheel \<br>	&#39;</pre><pre>docker_deploy_whl:<br>	docker exec -i $(CONTAINERNAME) bash -lc &#39;\<br>	  export PATH=&quot;$(EXEC_PATHS):$$PATH&quot;; \<br>	  cd &quot;$(APP_SUBDIR)&quot;; \<br>	  python -m twine upload -r local dist/*.whl \<br>	&#39;</pre><p>And boom 💥 — the exact same pipeline that was failing now builds and deploys successfully.</p><h3>🧰 Why This Works</h3><p>This solution:</p><ul><li>Doesn’t require modifying the base image</li><li>Doesn’t need source ~/.bashrc</li><li>Works identically in interactive and non-interactive shells</li><li>Keeps your pipeline clean and repeatable</li></ul><p>We’re simply ensuring the shell knows where poetry and twine live.</p><h3>🚀 Bonus Tip: Debugging PATH Inside a Container</h3><p>If you’re ever unsure where your tools are installed:</p><pre>docker exec -it &lt;container&gt; bash -lc &#39;command -v poetry&#39;<br>docker exec -it &lt;container&gt; bash -lc &#39;command -v twine&#39;</pre><p>Or run:</p><pre>docker exec -it &lt;container&gt; bash -lc &#39;ls -l /opt/venv/bin /root/.local/bin /usr/local/bin 2&gt;/dev/null&#39;</pre><h3>🧭 Final Thoughts</h3><p>If your build works manually inside a container but fails in your CI job, <strong>PATH is usually the culprit</strong>.</p><p>This little trick of exporting PATH explicitly has saved me (and now hopefully you) a ton of debugging time. It’s clean, generic, and works with any CI system — Jenkins, GitHub Actions, GitLab CI, or others.</p><h3>✅ TL;DR</h3><ul><li>Interactive shells set PATH, non-interactive don’t.</li><li>Poetry &amp; Twine live in /opt/venv/bin or similar.</li><li>Prepend those paths in your Makefile or pipeline.</li><li>Problem solved 🎉</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6a4f436b0c26" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[No Downtime, No Drama: Building Resilient Kubernetes Deployments]]></title>
            <link>https://medium.com/@satya.kumar/no-downtime-no-drama-building-resilient-kubernetes-deployments-d123b7fdf868?source=rss-749081e90d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/d123b7fdf868</guid>
            <dc:creator><![CDATA[Satya Kumar]]></dc:creator>
            <pubDate>Sun, 10 Aug 2025 17:12:21 GMT</pubDate>
            <atom:updated>2025-08-10T17:12:21.573Z</atom:updated>
            <content:encoded><![CDATA[<p><em>How to keep your apps running even when the unexpected happens</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cW8aa-EJ5zN6Akmi" /></figure><p><strong>Downtime is expensive.</strong><br> Every minute your application is offline can mean lost revenue, frustrated users, and a dent in your reputation.</p><p>Kubernetes gives us the tools to keep services online — but only if we configure them for <strong>high availability (HA)</strong>. In this guide, I’ll show you <strong>battle-tested techniques</strong> to make your workloads resilient and responsive, even during failures or upgrades.</p><h3>1. Run Multiple Pod Replicas</h3><p>Never rely on a single pod.</p><pre>spec:<br>  replicas: 2</pre><p>✅ <strong>Why it matters:</strong> If one pod fails or the node crashes, traffic shifts to other pods seamlessly.</p><h3>2. Spread Pods Across Nodes</h3><p>All replicas on one node = disaster waiting to happen.</p><p>Use <strong>Pod Anti-Affinity</strong> so pods run on different nodes (or zones):</p><pre>affinity:<br>  podAntiAffinity:<br>    requiredDuringSchedulingIgnoredDuringExecution:<br>    - labelSelector:<br>        matchExpressions:<br>        - key: app<br>          operator: In<br>          values:<br>          - my-app<br>      topologyKey: &quot;kubernetes.io/hostname&quot;</pre><p>📌 <strong>Result:</strong> A single node failure won’t take down your whole app.</p><h3>3. Deploy Without Downtime</h3><p>Use <strong>rolling updates</strong> to upgrade without stopping all pods at once:</p><pre>strategy:<br>  type: RollingUpdate<br>  rollingUpdate:<br>    maxSurge: 1<br>    maxUnavailable: 0</pre><p>💡 <strong>Tip:</strong> Pair this with readiness probes so only healthy pods get traffic.</p><h3>4. Shut Down Gracefully</h3><p>Give pods time to finish requests before termination:</p><pre>terminationGracePeriodSeconds: 60</pre><p>This avoids dropped connections during updates or scaling.</p><h3>5. Monitor Pod Health</h3><p>Kubernetes uses three types of probes:</p><p><strong>Probe Types &amp; Purposes</strong></p><ul><li><strong>Readiness</strong> — Send traffic only when the pod is ready</li><li><strong>Liveness</strong> — Restart pods that are unhealthy</li><li><strong>Startup</strong> — Allow slow-starting apps to initialize</li></ul><p>Example readiness probe:</p><pre>readinessProbe:<br>  httpGet:<br>    path: /health<br>    port: 8080</pre><h3>6. Right-Size Resources</h3><p>Prevent resource starvation by setting <strong>requests</strong> (min) and <strong>limits</strong> (max):</p><pre>resources:<br>  requests:<br>    cpu: &quot;250m&quot;<br>    memory: &quot;256Mi&quot;<br>  limits:<br>    cpu: &quot;500m&quot;<br>    memory: &quot;512Mi&quot;</pre><p>⚠ <strong>Without these</strong>, one noisy container can crash a node.</p><h3>7. Scale Intelligently</h3><p><strong>Horizontal Scaling (HPA):</strong></p><pre>kubectl autoscale deployment my-app --cpu-percent=70 --min=2 --max=10</pre><p><strong>Vertical Scaling:</strong> Adjust resources per pod dynamically.</p><p><strong>Cluster Scaling Tools:</strong></p><ul><li><strong>Cluster Autoscaler</strong> — scales node pools.</li><li><strong>Karpenter</strong> — faster, workload-aware scaling.</li></ul><h3>8. Protect Against Disruptions</h3><p>Use <strong>PodDisruptionBudgets (PDB)</strong> to ensure a minimum number of pods are always available:</p><pre>minAvailable: 2</pre><h3>Visual Overview</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iDUBY_IyPLqGikpU" /></figure><h3>Final Thoughts</h3><p>High availability isn’t magic — it’s <strong>intentional design</strong>. If you:</p><ul><li>Replicate workloads</li><li>Distribute pods smartly</li><li>Monitor health</li><li>Scale both pods and nodes</li></ul><p>…you can survive most outages without users noticing.</p><blockquote><em>🚀 </em><strong><em>Challenge:</em></strong><em> Test your setup. Deploy in staging and simulate a node failure:</em></blockquote><pre>kubectl drain &lt;node-name&gt; --ignore-daemonsets</pre><blockquote><em>Watch your app stay online.</em></blockquote><p>💬 <em>What’s your favorite Kubernetes HA practice? Share it in the comments.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d123b7fdf868" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Understanding Apache Druid Components — A Clear & Quick Guide]]></title>
            <link>https://medium.com/@satya.kumar/understanding-apache-druid-components-a-clear-quick-guide-d9aee28adba4?source=rss-749081e90d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/d9aee28adba4</guid>
            <category><![CDATA[data-pipeline]]></category>
            <category><![CDATA[druid]]></category>
            <category><![CDATA[apache-druid]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[data-platforms]]></category>
            <dc:creator><![CDATA[Satya Kumar]]></dc:creator>
            <pubDate>Wed, 30 Apr 2025 06:54:34 GMT</pubDate>
            <atom:updated>2025-04-30T06:54:34.025Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Understanding Apache Druid Components — A Clear &amp; Quick Guide</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vIoZjMrGujqnbkoDOTUgTw.png" /></figure><h3>Introduction</h3><p>Apache Druid is a real-time analytics database designed for <strong>ultra-fast query performance</strong> on massive datasets. But to a newcomer, its <strong>modular architecture</strong> can seem like a maze.</p><p>This blog simplifies Druid’s core components with a quick breakdown and a fun little story to help you <strong>visualize how everything works together</strong>.</p><h3>🧩 Druid Architecture — Core Components Explained</h3><p>Here’s a table summarizing what each component in Druid does🛠 How It All Works — In a Nutshell</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/7031b497ca568806cd4f46d903afd681/href">https://medium.com/media/7031b497ca568806cd4f46d903afd681/href</a></iframe><p>When data is ingested and queried in Druid, here’s what happens:</p><ol><li><strong>Overlord</strong> receives the ingestion request.</li><li>It assigns the job to a <strong>MiddleManager</strong>, which reads data and writes segments.</li><li>Once ready, segments are handed off to the <strong>Coordinator</strong>, which places them on <strong>Historical</strong> nodes.</li><li><strong>Broker</strong> receives user queries and asks the right Historical (or real-time) nodes for data.</li><li><strong>Router</strong> routes everything to the right service behind the scenes.</li></ol><h3>📖 A Simple Story: Druid as a Data Warehouse Company</h3><p>Imagine a high-tech warehouse company called <strong>Druid Inc.</strong>. Here’s how the team operates:</p><ul><li>🧠 <strong>Coordinator</strong> is the warehouse manager. They decide where boxes (data segments) should be stored and ensure backups exist.</li><li>🛠 <strong>Overlord</strong> is the task manager. They assign workers to process incoming deliveries.</li><li>⚙️ <strong>MiddleManager</strong> is the worker. They unpack raw deliveries (raw data), clean it, label it, and store it properly.</li><li>🧳 <strong>Historical</strong> nodes are storage rooms, holding the clean boxes and ready to serve them.</li><li>🚪 <strong>Broker</strong> is the receptionist. You ask them a question (“How many shoes sold yesterday?”) — and they know exactly which storage rooms to check.</li><li>🌐 <strong>Router</strong> is the front desk. It knows which department to send your request to.</li><li>📊 <strong>Metrics Exporter</strong> is the dashboard on the manager’s wall — showing how everyone’s doing.</li></ul><p>Together, Druid Inc. runs like a well-oiled machine — powering high-speed analytics for its customers.</p><h3>🔚 Wrapping Up</h3><p>Druid’s architecture might look complex at first, but once you understand the roles of each component, it becomes easy to grasp. Just remember the story of <strong>Druid Inc.</strong> — every piece has a purpose, and they all work together to deliver <strong>real-time insights</strong> at lightning speed.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d9aee28adba4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How I Recovered 287GB of Disk Space by Fixing Open Deleted Files on Docker and LocalStack]]></title>
            <link>https://medium.com/@satya.kumar/how-i-recovered-287gb-of-disk-space-by-fixing-open-deleted-files-on-docker-and-localstack-6dfc7e63ac15?source=rss-749081e90d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/6dfc7e63ac15</guid>
            <category><![CDATA[disk-space-management]]></category>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[file-handling]]></category>
            <category><![CDATA[troubleshooting]]></category>
            <dc:creator><![CDATA[Satya Kumar]]></dc:creator>
            <pubDate>Sat, 08 Feb 2025 13:10:48 GMT</pubDate>
            <atom:updated>2025-02-08T13:13:36.484Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/635/1*tZbI5DgyuYiJ0d7SmDda-Q.png" /></figure><p><strong>How I Recovered 287GB of Disk Space by Fixing Open Deleted Files on Docker and LocalStack</strong></p><p><strong>Introduction</strong></p><p>Recently, while working on a <strong>Docker-based project with LocalStack</strong>, I noticed that my server’s disk space usage was growing rapidly. Upon running basic disk checks, I discovered a huge discrepancy between what was reported as <strong>used disk space</strong> and the <strong>actual files visible in the system</strong>. After digging deeper, I found the issue was related to <strong>open deleted files being held by running processes</strong>.</p><p>However, upon further investigation, I realized that the <strong>main culprit</strong> was my initial setup of the LocalStack container <strong>without properly mounting volumes</strong>. When volumes were not specified, many temporary files generated inside the container were tied to the container’s internal filesystem. These files were accumulating and couldn’t be properly freed, even after deletion, because they were still referenced by running processes.</p><p>Once I <strong>added persistent volumes</strong> to the LocalStack container, the problem of recurring storage usage drastically reduced, and the disk space issues disappeared. Although this adjustment solved the core issue, the troubleshooting process remains valuable for debugging and understanding <strong>how open deleted files</strong> or hidden system references can cause disk space discrepancies.</p><p>In this blog, I’ll walk you through <strong>how I debugged, identified, and resolved the issue</strong> — ultimately freeing up <strong>287GB of disk space</strong> in the process! If you’re facing similar disk space concerns, these troubleshooting tips should help you debug effectively and resolve the problem.</p><p><strong>The Problem: Missing Disk Space</strong></p><p>Initially, I ran the following commands to check disk usage:</p><pre># Check overall disk usage in the /data directory<br>du -sh /data/<br>64G    /data/<br><br># Check disk usage reported by the filesystem<br>df -h /data<br>Filesystem                        Size  Used Avail Use% Mounted on  <br>/dev/mapper/ubuntu--vg-data--lv   367G  311G   38G  90% /data  </pre><p><strong>Problem:</strong> Although the du command reported only 64GB of data in /data, the df command showed that 311GB of space was being used. I was missing nearly 247GB of disk space!</p><p><strong>Initial Investigation</strong></p><p>To find where the missing space was going, I suspected deleted files being held open by running processes could be the cause. These files are not visible in the filesystem but still consume disk space until the processes holding them are terminated.</p><p>I ran the following command to check for open deleted files:</p><p>lsof | grep &#39;(deleted)</p><p>I found thousands of entries like this:</p><pre>python 1558065 root 202u REG 0,67 744225 /tmp/localstack-s3-storage/hyphen/tmpx1s29y__ (deleted)<br>python 1558065 root 203u REG 0,67 745436 /tmp/localstack-s3-storage/hyphen/tmpcchlk8ww (deleted)</pre><p>…</p><p>These were temporary files generated by LocalStack’s S3 emulation service, stored in /tmp/localstack-s3-storage/hyphen/. Although they had been deleted, they were still consuming space because they were being held open by running Python processes.</p><p><strong>Step-by-Step Solution</strong></p><p><strong>Step 1: Identify and Kill Processes Holding Deleted Files</strong></p><p>To fix this, I needed to terminate the processes keeping these files open.</p><p>First, I listed the unique process IDs (PIDs) holding the deleted files:</p><blockquote>lsof | grep ‘(deleted)’ | awk ‘{print $2}’ | sort -u</blockquote><p>Then, I killed all the processes in bulk using:</p><blockquote>lsof | grep ‘(deleted)’ | awk ‘{print $2}’ | sort -u | xargs -r kill -9</blockquote><p>This command finds all unique PIDs and forcefully terminates them. Be careful when using kill -9 because it can terminate critical services. Ensure you know which processes you are killing.</p><p><strong>Step 2: Verify Disk Space</strong></p><p>After killing the processes, I rechecked the disk usage:</p><blockquote>df -h /data</blockquote><p><strong>Before cleanup:</strong></p><p>Filesystem Size Used Avail Use% Mounted on</p><blockquote>/dev/mapper/ubuntu — vg-data — lv 367G 311G 38G 90% /data</blockquote><p><strong>After cleanup:</strong></p><p>Filesystem Size Used Avail Use% Mounted on</p><blockquote>/dev/mapper/ubuntu — vg-data — lv 367G 24G 325G 7% /data</blockquote><p><strong>Success!</strong> I had freed up 287GB of disk space.</p><p><strong>Why Did This Happen?</strong></p><p>The issue occurred because LocalStack’s S3 emulation service was creating temporary files that were deleted but not properly closed by running processes. As a result, the files continued consuming disk space until the processes were terminated.</p><p>This is a common issue in environments using:</p><p>• Docker containers</p><p>• Services with large temporary file usage</p><p>• Log files that grow and are frequently rotated</p><p><strong>Step 3: Preventing Future Issues</strong></p><p><strong>1. Configure LocalStack to Use a Persistent Storage Location</strong></p><p>Instead of using /tmp, I configured LocalStack to use a more controlled storage directory:</p><p><strong>2. Set Up a Cron Job to Clean Temporary Files Periodically</strong></p><p>I created a cron job to clean up old files:</p><blockquote>sudo crontab -e</blockquote><p>I added the following line to delete files older than 1 day in /tmp:</p><blockquote>0 0 * * * find /tmp/localstack-s3-storage/ -type f -mtime +1 -delete</blockquote><p><strong>3. Regularly Clean Up Docker Resources</strong></p><p>Docker volumes, images, and containers can also contribute to disk usage. I set up a regular cleanup:</p><blockquote><em># Clean unused Docker resources<br></em>sudo docker system prune -a — volumes -f</blockquote><p><strong>Conclusion</strong></p><p>By identifying open deleted files and cleaning them up, I successfully reclaimed 287GB of disk space. The key takeaway here is that deleted files aren’t truly gone until the processes using them are terminated. If you encounter similar issues, follow these steps to debug and free up disk space.</p><p><strong>Key Commands Summary</strong></p><pre># Find processes holding deleted files<br>lsof | grep &#39;(deleted)&#39; <br># Kill all processes holding deleted files<br>lsof | grep &#39;(deleted)&#39; | awk &#39;{print $2}&#39; | sort -u | xargs -r kill -9<br># Check disk usage<br>df -h /data<br># Clean unused Docker resources<br>sudo docker system prune -a - volumes -f</pre><p>I hope this blog helps you debug disk space issues effectively. Let me know in the comments if you have encountered similar issues or if you have any tips to share!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6dfc7e63ac15" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Learning Path for DevOps]]></title>
            <link>https://medium.com/@satya.kumar/learning-path-for-devops-c6e66e5f1f01?source=rss-749081e90d99------2</link>
            <guid isPermaLink="false">https://medium.com/p/c6e66e5f1f01</guid>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[learning]]></category>
            <category><![CDATA[continuous-integration]]></category>
            <dc:creator><![CDATA[Satya Kumar]]></dc:creator>
            <pubDate>Tue, 04 Jan 2022 12:08:15 GMT</pubDate>
            <atom:updated>2022-01-25T05:34:29.036Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*tyFbJtARBrjRVH4-49pFCw.png" /></figure><h3>What is DevOps</h3><p>Simply put, DevOps is “how you build and deploy your software.”</p><p>It’s about your team’s culture, the ideas that guide your development process, the set of techniques you use to automate and speed up your delivery, and the mindset you exhibit to achieve complete client satisfaction.</p><p>DevOps is a shortened acronym for Development and Operations. In essence, it is bringing your software development and operations teams together to work toward a common goal and share a common vision. Rather than individual performances or targets, their collective attention is always on the product or application that is being developed.</p><h3>1. Start by Learning GIT</h3><p>GIT is one of the most widely used versioning systems in the IT industry. You don’t need to be an expert in GIT to use it, but it is a technology that will accompany you throughout your DevOps journey.</p><p>Check this video of <a href="https://www.youtube.com/channel/UCBGOUQHNNtNGcGzVq5rIXjw">Kunal Kushwaha</a> on a tutorial for Git and Github. This is an excellent lesson for learning all there is to know about Github.</p><p>Link: <a href="https://youtu.be/apGV9Kg7ics">https://youtu.be/iwolPf6kN-k</a></p><blockquote>Here are few important GIT commands.</blockquote><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/922250f04551971a1dcf1d61d87bbd31/href">https://medium.com/media/922250f04551971a1dcf1d61d87bbd31/href</a></iframe><h3><strong>2. Learn A Programming Language</strong></h3><p>Programming skills are nice to have. A good DevOps Engineer must have good knowledge on Development and Operations.</p><blockquote>Python, Go, JavaScript and few among popular programming languages that add value to the DevOps Career.</blockquote><h3>3. Learn Server Management (LINUX,LINUX,…)</h3><p>Managing servers is one of the main responsibilities of a DevOps practitioner. Knowing how servers work is essential, and to do so, some understanding of hardware (CPU, architecture, memory, etc.) is required. Another thing to learn is about operating systems, particularly Linux. You might begin by selecting a distribution such as Ubuntu.</p><blockquote>Learning Application and Network Troubleshooting commands is a must. Learn Basics of Linux Administration ( RHCSA is a good path ) &amp; Apply that in your daily work! Focus on process management, file management, Systemd, Log rotation.Below are few basic Linux Commands.</blockquote><p>This is a must watch video of <a href="https://www.youtube.com/channel/UCBGOUQHNNtNGcGzVq5rIXjw">Kunal Kushwaha</a> on some of the most important Linux commands.</p><p>Link: <a href="https://youtu.be/iwolPf6kN-k">https://youtu.be/iwolPf6kN-k</a></p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/c7151b2d2d3b0484329d51ec1f58aaa7/href">https://medium.com/media/c7151b2d2d3b0484329d51ec1f58aaa7/href</a></iframe><h3>4. Learn Networking &amp; Security Basics</h3><p>Knowledge of networking is very important in DevOps; the basic notion of networking is very important in DevOps since it allows for continuous integration and connectivity with other servers, which is necessary for continuous development.</p><blockquote>One should understand how HTTP, DNS, FTP, and other protocols function, as well as how to secure your deployed software, anticipate security issues in the code, and configure your infrastructure network.</blockquote><h3>5. Learn Scripting</h3><p>Shell programming or bash scripting is required knowledge for aspiring DevOps engineers. One can focus on both Bash and Python scripting which widely helps in automation</p><blockquote>Here are a few of the reasons I utilise scripts.</blockquote><blockquote>1. You can develop a script to initialise something during the system’s boot process. As a result, you won’t have to do anything manually.<br>2. You can design a script that instals each pre-requisite separately and then builds the code with user input to enable/disable specific features.<br>3. To simultaneously kill or start numerous apps.<br>4. To look over a vast database of files and see if there are any patterns.<br>So, in general, to automate the procedure, and so on….</blockquote><h3>6. Learn How to Install &amp; Configure Middle-wares</h3><p>In the DevOps world, Apache and Nginx are the most often used middleware, and knowing how to install and configure virtual hosts, reverse proxies, domain names, and SSL can greatly aid you in your everyday responsibilities.</p><blockquote>Start by deploying Nginx as a web server for a simple microservice based application</blockquote><h3>7. Learn How To Build an Application</h3><p>Building of application comes before running. Building software is generally about running a procedure of creating a software release that could run in a production server. DevOps professionals need to know about this important part of the software lifecycle.</p><blockquote><em>Create an application in the language of your choice and check the different ways to install its dependencies and build its code.</em></blockquote><h3>8. Learn How To Deploy an Applications</h3><p>You’ll need to know how to deploy apps to a production server once you’ve learned how to install and configure Nginx.</p><blockquote><em>Create a “hello world” applications using Python, Nodejs and Java. Deploy these 3 applications. You can use Nginx as a reverse proxy for all of them.</em></blockquote><h3>9. Learn How To Automate Your Application deployment</h3><p>DevOps is not solely automation. Automation is absolutely a huge part of DevOps. Once you learned how to build software, you can use tools like Jenkins to automate builds and connect your code to the code repository.</p><blockquote>Read about Continuous Integration and Continuous Delivery if you’re not familiar.</blockquote><blockquote>- Install Jenkins<br>- Configure a declarative pipeline using Jenkinsfile inside the git repo of your app<br>- Implement 2 stages: build , deploy.<br>- Embed other stages to establish confidence : run unit-tests before build stage.. run static code analysis before the build stage also.<br>- Deploy stage requires a communication protocol between CI/CD system and target deployment(s): If it’s VM based deployment, SSH is the protocol. If it’s Kubernetes-based deployment, HTTP protocol (REsT of kubeapi) is the protocol.<br>Hint: Deploying containers without kubernetes is classified as VM-based deployment and SSH is required.</blockquote><h3>10. Learn any of the Public Cloud</h3><p>There are multiple cloud providers available and I suggest to start with AWS, GCP or Azure. One can also focus on beginner level certification which helps in understanding services and use case scenarios.</p><p>This also gets you started on the Cloud by giving you an overview of the various cloud services and how they link, as well as easing you into the realm of Infra/Ops (Resilience, Failure, and DR, Loadbalancing, Scalability, Subnetting, Supernetting, Routing, Firewalls, WAF, and so on).</p><blockquote>Read about Continuous Integration and Continuous Delivery if you’re not familiar.</blockquote><h3>11. Learn a Configuration Management tool</h3><p>Learning a configuration management tool will make your life easier as things become more complex and you need to handle different environments and configurations.</p><p>There are several configuration management solutions available, such as Saltstack, Ansible, Chef, Puppet, and so on, and you may find a resource that compares these tools online. Choose a configuration management tool based on your requirements and begin learning it.</p><blockquote>Once you learn how to build and deploy applications, you can write a playbook that installs all dependencies &amp; Deploy the application on multiple servers.</blockquote><h3>12. Learn Infrastructure As Code</h3><p>IaC is absolutely important to automate your infrastructure and provision your environments with simple scripts or alternative tools. Apply Pipelines not only for your Apps but also for its Infra.</p><p>Infra as Code should make you think about infrastructural upgrades as software. It must have unit tests, integration tests, and static analysis because it is software. Use the correct technologies/tools (e.g., Terraform, CloudFormation,…) that include key infrastructure functionalities as code ( modularity, reusability, artifacting, externalising secret data smoothly).</p><blockquote>Choose a cloud provider (AWS, GCP ..etc) and you will find a lot of free online resources to start your infrastructure. You can also learn how to use “cloud managers” technologies, some Configuration Management tools like Saltstack could help you provision infrastructure on AWS or GCP, otherwise, if you need more go for technologies like Terraform.</blockquote><h3>13. Learn About Containers &amp; Orchestration</h3><p>Containers, such as Docker, are quickly becoming a must-have talent. You’ll need to know how to build, deploy, and manage containers in both development and production contexts.</p><p>Container orchestration (Kubernetes) is used by software teams to handle and automate a variety of tasks, including container provisioning and deployment, container redundancy and availability, and container management. Increasing or decreasing the number of containers to distribute application load uniformly across the host infrastructure</p><blockquote>- Learn the fundamentals of containers from two perspectives: SysOps (linux processes) and Developers (app package)<br>- Learn Docker first, then Docker Compose. Dockerize the app(s) you created in the 7th and 8th parts. (Dockerfile, etc.)<br>- Create an AWS EC2 instance or a Google Compute Engine instance. Install Docker Dependencies. Deploy your application (s).<br>- Learn about the Kubernetes architecture, how it works, and how the entire infrastructure is exposed via a REST API!!<br>- Deploy your application’s container image to a public registry (Docker hub).<br>- In order to deploy the app(s) mentioned earlier, create the necessary Kubernetes resources in YAML format.<br>- Using kubectl, deploy your app.<br>- Scale up your app</blockquote><h3>14. Learn How To Monitor Application &amp; Infrastructure</h3><p>A application that has been deployed in production, as well as the infrastructure that supports it, should be monitored. One of the most critical abilities you’ll need is monitoring and alerting.</p><blockquote>To get started, open source tools such as Prometheus, Grafana, and ELK Stack are available. You might also want to learn how to set up and use an ELK stack.</blockquote><h3>Never Stop Learning</h3><p>To survive in the DevOps realm, we must constantly update our knowledge of Cloud and DevOps tools and techniques. We must stay current with the trend and continue to learn.<br>After you’ve completed all of the preceding tasks, you can begin focusing on concepts such as Service Mesh, Chaos Engineering, Deployment Strategies, Branching Strategies, Helm, and so on.</p><blockquote><strong>Try to put everything you’ve learned so far into practise, and come up with new ways to introduce security and best practises along the way.</strong></blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c6e66e5f1f01" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>