<?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 Suraj Solanki on Medium]]></title>
        <description><![CDATA[Stories by Suraj Solanki on Medium]]></description>
        <link>https://medium.com/@surajblog?source=rss-b6f5f0f84810------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*2EHz-ACI1Zhdt5NQSBwtoQ.jpeg</url>
            <title>Stories by Suraj Solanki on Medium</title>
            <link>https://medium.com/@surajblog?source=rss-b6f5f0f84810------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 09:37:13 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@surajblog/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[Integrating Kasten K10 with TrueNAS for Backup on OpenShift]]></title>
            <link>https://surajblog.medium.com/integrating-kasten-k10-with-truenas-for-backup-on-openshift-5a302d756136?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/5a302d756136</guid>
            <category><![CDATA[truenas]]></category>
            <category><![CDATA[kasten]]></category>
            <category><![CDATA[disaster-recovery]]></category>
            <category><![CDATA[openshift]]></category>
            <category><![CDATA[backup]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Sat, 11 Apr 2026 16:02:05 GMT</pubDate>
            <atom:updated>2026-04-11T16:02:05.982Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c9-kiSrKbarYe5ggeXLFTA.png" /></figure><p>Modern Kubernetes environments require reliable backup and disaster recovery. <strong>Kasten K10</strong> is a Kubernetes-native data protection platform that provides policy-driven backups, application mobility, and disaster recovery. When combined with <strong>TrueNAS NFS storage</strong>, it becomes a powerful and cost-effective backup repository for OpenShift workloads.</p><p>This article explains how to integrate Kasten K10 with TrueNAS using NFS storage to enable automated application backups.</p><h3>Prerequisites</h3><p>Before you begin, ensure the following are in place:</p><ul><li><strong>TrueNAS Server : </strong>A TrueNAS server with the NFS service running</li><li><strong>TrueNAS NFS Host : </strong>IP address or hostname of the TrueNAS server</li><li><strong>NFS Mount Path : </strong>Dataset path exported via NFS</li><li><strong>Storage Capacity : </strong>Sufficient storage allocated for backups</li><li><strong>NFS Permissions : A</strong>ppropriate dataset permissions configured</li><li><strong>Kasten K10 : </strong>Installed on the OpenShift cluster in the kasten-io namespace</li><li><strong>VolumeSnapshotClass</strong></li></ul><h3>Why Use TrueNAS as a Backup Target</h3><p>TrueNAS provides enterprise-grade storage with strong data integrity.</p><p>Key advantages:</p><p><strong>ZFS filesystem</strong></p><ul><li>End-to-end checksums</li><li>Data corruption protection</li><li>Snapshot capability</li></ul><p><strong>NFS support</strong></p><ul><li>Native integration with Kubernetes</li><li>RWX access mode</li><li>Easy external storage integration</li></ul><p><strong>Large backup repository</strong></p><ul><li>Ideal for long-term backup retention</li><li>Cost-efficient compared to cloud object storage</li></ul><h3>Creating the TrueNAS NFS Backup Repository</h3><p>First, create a dataset inside the storage pool.</p><p>Example dataset:</p><pre>/mnt/pool0/k10-repo</pre><p>Then configure the NFS share.</p><p>Important settings:</p><pre>| Setting       | Value      |<br>| ------------- | ---------- |<br>| Maproot User  | root       |<br>| Maproot Group | wheel      |</pre><p>Disabling <strong>root squashing</strong> ensures Kasten can write backup data.</p><h3>Connecting OpenShift to TrueNAS</h3><p>Create a static <strong>PersistentVolume</strong> pointing to the TrueNAS NFS share.</p><pre>kind: PersistentVolume<br>apiVersion: v1<br>metadata:<br>  name: truenas-persistent-volume<br>spec:<br>  capacity:<br>    storage: 50Gi<br>  nfs:<br>    server: 112.12.11.21 # Your TrueNAS server IP<br>    path: /mnt/pool0/k10-repo # NFS export path on TrueNAS<br>  accessModes:<br>    - ReadWriteMany<br>  mountOptions:<br>    - hard<br>    - nfsvers=4.1<br><br><br><br>Expected output:<br><br>```<br>NAME                        STATUS   VOLUME       CAPACITY   ACCESS MODES   STORAGECLASS   <br>truenas-persistent-volume   Bound    truenas-pv     50Gi       RWX            nfs            <br>```</pre><p>Then create a <strong>PVC in the kasten namespace</strong>.</p><pre>kind: PersistentVolumeClaim<br>apiVersion: v1<br>metadata:<br>  name: truenas-persistent-volume-claim<br>  namespace: kasten-io<br>spec:<br>  accessModes:<br>    - ReadWriteMany<br>  resources:<br>    requests:<br>      storage: 50Gi<br>  volumeName: truenas-persistent-volume<br>  storageClassName: &quot;&quot; #it should be nfs storage class or cephfs<br>  volumeMode: Filesystem</pre><p>This PVC becomes the <strong>backup repository for Kasten exports</strong>.</p><h3>Creating the Kasten Location Profile</h3><p>Kasten uses a <strong>Location Profile</strong> to define backup storage.</p><pre>kind: Profile<br>apiVersion: config.kio.kasten.io/v1alpha1<br>metadata:<br>  name: truenas-profile<br>  namespace: kasten-io<br>spec:<br>  type: Location<br>  locationSpec:<br>    type: FileStore<br>    fileStore:<br>      claimName: truenas-persistent-volume-claim<br>      path: /</pre><p>This configuration tells Kasten to store backup exports on the <strong>TrueNAS NFS volume</strong>.</p><h3>Create the Shallow StorageClass</h3><p>If your cluster uses ODF/OCS with CephFS, create a shallow StorageClass to optimize export performance.</p><p>Without this:</p><ul><li>Full clone is created</li><li>Higher storage usage</li><li>Slower backups</li></ul><p>With backingSnapshot: &quot;true&quot;:</p><ul><li>Instant snapshot-backed clone</li><li>Minimal storage usage</li><li>Faster exports</li></ul><pre>kind: StorageClass<br>apiVersion: storage.k8s.io/v1<br>metadata:<br>  name: cephfs-shallow<br>  annotations:<br>    k10.kasten.io/sc-supports-block-mode-exports: &#39;true&#39;<br>    k10.kasten.io/sc-supports-read-only-mount: &#39;true&#39;<br>provisioner: openshift-storage.cephfs.csi.ceph.com<br>parameters:<br>  clusterID: openshift-storage<br>  fsName: ocs-storagecluster-cephfilesystem<br>  backingSnapshot: &#39;true&#39;<br>reclaimPolicy: Delete<br>allowVolumeExpansion: true<br>volumeBindingMode: Immediate</pre><h3>Annotate the VolumeSnapshotClass</h3><p>Kasten requires a VolumeSnapshotClass annotated with its specific annotation. Without this, backup policies will fail with a “Failed to find VolumeSnapshotClass” error.</p><pre># Find your existing VolumeSnapshotClasses:<br><br><br>$ oc get volumesnapshotclass<br><br><br># Annotate the one that matches your storage provisioner<br><br># For CephFS<br><br>$ oc annotate volumesnapshotclass ocs-storagecluster-cephfsplugin-snapclass \<br>  k10.kasten.io/is-snapshot-class=true<br><br># For RBD<br>$ oc annotate volumesnapshotclass ocs-storagecluster-rbdplugin-snapclass \<br>  k10.kasten.io/is-snapshot-class=true<br></pre><h3>Backup Policy Example</h3><p>Kasten policies define how often backups run.</p><p>Example weekly policy:</p><pre>apiVersion: config.kio.kasten.io/v1alpha1<br>kind: Policy<br>metadata:<br>  name: true-nas-snapshot<br>  namespace: kasten-io<br>spec:<br>  frequency: &quot;@weekly&quot;<br>  retention:<br>    daily: 7<br>  selector:<br>    matchExpressions:<br>      - key: k10.kasten.io/appNamespace<br>        operator: In<br>        values:<br>          - production      # Replace with your app namespace<br>  actions:<br>    - action: backup<br>      backupParameters:<br>        filters: {}<br>        profile:<br>          name: truenas-profile<br>          namespace: kasten-io<br>        imageRepoProfile:<br>          name: truenas-profile<br>          namespace: kasten-io<br>    - action: export<br>      exportParameters:<br>        frequency: &quot;@daily&quot;<br>        profile:<br>          name: truenas-profile<br>          namespace: kasten-io<br>        exportData:<br>          enabled: true<br>          overrides:<br>            - storageClassName: ocs-storagecluster-cephfs<br>              enabled: true<br>              exporterStorageClassName: cephfs-shallow<br>      retention:<br>        daily: 30</pre><h3>Verify the Setup</h3><pre># Check all resources<br><br><br>$ oc get pvc truenas-pvc -n kasten-io <br><br>$ oc get profile true-nas -n kasten-io <br><br>$ oc get policy true-nas-dailysnapshot -n kasten-io<br><br>$ oc get volumesnapshotclass<br><br><br># Trigger a manual policy run<br><br><br>$ oc -n kasten-io annotate policy true-nas-dailysnapshot \<br>  k10.kasten.io/manualTrigger=$(date +%s) --overwrite<br><br># Monitor the run<br><br>$ oc get actions -n kasten-io --sort-by=.metadata.creationTimestamp | tail -5</pre><h3>Conclusion</h3><p>Integrating Kasten K10 with TrueNAS provides a <strong>cost-effective, enterprise-grade backup solution</strong> for OpenShift workloads.</p><p>With this setup:</p><ul><li>Fast restores are handled via CSI snapshots</li><li>Long-term retention is managed through TrueNAS</li><li>Backup operations are automated and scalable</li></ul><p>This combination ensures both <strong>operational efficiency and disaster recovery readiness</strong> for Kubernetes environments.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5a302d756136" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kasten K10 Disaster Recovery (DR) Setup Guide for Kubernetes / OpenShift]]></title>
            <link>https://surajblog.medium.com/kasten-k10-disaster-recovery-dr-setup-guide-for-kubernetes-openshift-3a36a360384c?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/3a36a360384c</guid>
            <category><![CDATA[veeam-backup]]></category>
            <category><![CDATA[veeam-backup-software]]></category>
            <category><![CDATA[kubernetes-cluster]]></category>
            <category><![CDATA[openshift]]></category>
            <category><![CDATA[disaster-recovery]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Mon, 16 Mar 2026 12:18:45 GMT</pubDate>
            <atom:updated>2026-03-16T12:18:45.586Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aZq9syWNVE8NY0980YsAxQ.png" /></figure><p>Modern Kubernetes platforms require strong <strong>disaster recovery (DR)</strong> strategies. While application backups protect workloads and persistent volumes, it is equally important to protect the <strong>backup system itself</strong>.</p><p>In <strong>Kasten K10</strong>, the <strong>catalog</strong> stores critical backup metadata such as:</p><ul><li>Backup policies</li><li>Restore points</li><li>Storage location profiles</li><li>Backup configuration and metadata</li></ul><p>If the cluster fails and the catalog is lost, restoring applications becomes difficult even if backup data still exists.</p><p>To solve this, <strong>Kasten K10 Disaster Recovery (DR)</strong> allows you to export the K10 catalog to <strong>external object storage</strong> such as:</p><ul><li>AWS S3/Azure Storage/ Google Cloud Storage</li><li>NFS/SMB</li><li>Other S3-compatible object storage systems</li></ul><p>This guide explains how to configure <strong>Kasten K10 Disaster Recovery using the K10 Dashboard UI on OpenShift</strong>.</p><h3>Why Kasten DR Is Important</h3><p>In Kubernetes environments, clusters can fail due to:</p><ul><li>Infrastructure failures</li><li>Storage corruption</li><li>Accidental deletion of namespaces</li><li>Cluster rebuild or migration</li></ul><p>Without protecting the <strong>Kasten catalog</strong>, you may lose:</p><ul><li>Backup policies</li><li>Restore history</li><li>Backup configuration</li></ul><p>Kasten DR ensures that <strong>all metadata required for recovery is safely stored outside the cluster</strong>.</p><h3>Prerequisites</h3><p>Before configuring disaster recovery, ensure the following requirements are met.</p><p>Kasten K10 must be installed and running on your <strong>OpenShift cluster</strong>.</p><p>You need an object storage bucket for DR backups, such as:</p><ul><li>AWS S3/Azure Storage/ Google Cloud Storage</li><li>NFS/SMB</li><li>Any S3-compatible storage system</li></ul><p>You must have the following credentials:</p><ul><li><strong>Access Key ID</strong></li><li><strong>Secret Access Key</strong></li></ul><h3>Step 2 — Kasten Dashboard Access</h3><p>The <strong>K10 Dashboard</strong> must be accessible via browser.</p><p>Example route:</p><pre>https://k10-route-kasten-io.apps.&lt;cluster-domain&gt;/k10/#</pre><p>Log in using your <strong>authentication credentials</strong>.</p><h3>Step 2 — Create an S3 Location Profile</h3><p>The <strong>Location Profile</strong> defines where Kasten stores backup data and disaster recovery metadata.</p><p>Navigate to console : <strong>Settings → Profiles → Location → New Profile</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/793/1*0nh4VvCHOemJld8YglXsFA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/725/1*5bd53yZ_aVNcpSwSsY48eQ.png" /></figure><p><strong>Immutable Backups : </strong>Enable this option if supported by your object storage.<br> This prevents backups from being modified or deleted.</p><p>After entering the configuration:</p><ol><li>Click <strong>Validate</strong></li><li>Click <strong>Save</strong></li></ol><p>Wait for the validation status to show: SUCCESS</p><p>This confirms that Kasten can access the storage bucket.</p><ol><li><strong>Configure the profile with the following values using cli.</strong></li></ol><pre>kind: Profile<br>apiVersion: config.kio.kasten.io/v1alpha1<br>metadata:<br>  name: profile-name<br>  namespace: kasten-io<br>spec:<br>  locationSpec:<br>    type: ObjectStore<br>    objectStore:<br>      name: s3-bucket-name<br>      objectStoreType: S3<br>      region: us-east-1<br>    credential:<br>      secretType: AwsAccessKey<br>      secret:<br>        apiVersion: v1<br>        kind: secret<br>        name: k10secret-key # create first secret for aws access key and secret key<br>        namespace: kasten-io<br>  type: Location</pre><h3>Step 3 — Enable Kasten Disaster Recovery</h3><p>Once the location profile is created, enable DR for the Kasten catalog.</p><p>Navigate to console : Settings → Setup Kasten DR</p><p><strong>Configure the following options:</strong></p><h4>Location Profile</h4><pre>Select the S3 profile you created earlier</pre><h4>Passphrase Method</h4><pre>Passphrase</pre><p>This passphrase is used to <strong>encrypt catalog backups</strong>.</p><h4>Snapshot Configuration</h4><pre>Export local catalog snapshots</pre><p>Click <strong>Enable Kasten DR</strong> to activate disaster recovery.</p><h3>What Happens Internally</h3><p>Once DR is enabled, Kasten automatically creates a <strong>disaster recovery policy</strong> responsible for exporting the catalog to object storage.</p><p>Example policy generated by Kasten:</p><pre>apiVersion: config.kio.kasten.io/v1alpha1<br>kind: Policy<br>metadata:<br>  name: k10-disaster-recovery-policy<br>  namespace: kasten-io<br>spec:<br>  frequency: &quot;@hourly&quot;<br>  retention:<br>    hourly: 4<br>    daily: 1<br>    weekly: 1</pre><p>This policy performs the following tasks automatically:</p><ul><li>Creates <strong>hourly snapshots of the Kasten catalog</strong></li><li><strong>Encrypts the metadata</strong> using the configured passphrase</li><li><strong>Exports the encrypted snapshots to S3 object storage</strong></li></ul><p>This ensures that the Kasten configuration is always recoverable.</p><h3>Disaster Recovery Data Flow</h3><p>The DR workflow follows this process:</p><ol><li><strong>Kasten Catalog</strong> stores policies, profiles, and backup metadata.</li><li>Kasten creates a <strong>catalog snapshot</strong>.</li><li>The snapshot is <strong>encrypted using the passphrase</strong>.</li><li>An <strong>export job</strong> is triggered.</li><li>The encrypted snapshot is uploaded to the <strong>S3 bucket</strong>.</li></ol><p>The exported catalog includes:</p><ul><li>Backup policies</li><li>Location profiles</li><li>Restore points metadata</li><li>Backup configuration</li></ul><h3>Verify the DR Configuration</h3><p>You can verify the disaster recovery configuration using the <strong>OpenShift CLI</strong>.</p><h4>Check the DR Policy</h4><pre>oc get policy k10-disaster-recovery-policy -n kasten-io</pre><h4>Check Recent Actions</h4><pre>oc get actions -n kasten-io</pre><p>This shows recent backup or export jobs.</p><h4>Verify Encryption Key Secret</h4><pre>oc get secret k10masterkey -n kasten-io</pre><p>If everything is working correctly, you will see <strong>scheduled DR export jobs running periodically</strong>.</p><h3>Disaster Recovery Restore Process</h3><p>In case of a <strong>complete cluster failure</strong>, the recovery process is simple.</p><ol><li>Install <strong>Kasten K10</strong> on the new cluster.</li><li>Recreate the <strong>same S3 Location Profile</strong>.</li><li>Navigate to <strong>Setup Kasten DR</strong>.</li><li>Enter the <strong>original passphrase</strong>.</li><li>Provide the <strong>Cluster ID</strong>.</li></ol><p>Kasten will automatically restore:</p><ul><li>Backup policies</li><li>Location profiles</li><li>Restore points metadata</li></ul><p>This restores the <strong>entire backup configuration environment</strong>.</p><h3>Important Things to Remember</h3><h4>Save the Passphrase</h4><p>The passphrase is required to decrypt DR backups.</p><p>If the passphrase is lost, the DR backups <strong>cannot be restored</strong>.</p><p>Store the passphrase securely using:</p><ul><li>Password managers</li><li>Secret management tools</li><li>Secure documentation</li></ul><h4>Save the Cluster ID</h4><p>The <strong>Cluster ID</strong> is required when restoring the Kasten catalog.</p><p>Without the Cluster ID, Kasten cannot identify the correct catalog backups.</p><h4>DR Protects Only Kasten Metadata</h4><p>Kasten DR protects <strong>backup metadata only</strong>.</p><p>Application data must still be protected using <strong>regular backup policies</strong>.</p><h3>Best Practice Architecture</h3><h4>Layer 1 — Application Backup</h4><p>Application data and persistent volumes are backed up to <strong>TrueNAS NFS storage</strong>.</p><h4>Layer 2 — Kasten DR Metadata</h4><p>Kasten catalog metadata is exported to <strong>S3 object storage</strong>.</p><p>This design provides <strong>two independent recovery layers</strong>, ensuring that both application data and backup configuration remain protected.</p><h3>Conclusion</h3><p>Kasten K10 Disaster Recovery ensures that the <strong>backup platform itself is protected</strong>. By exporting the <strong>K10 catalog to S3 object storage</strong>, organizations can recover backup policies, metadata, and restore points even after a complete cluster failure. When combined with <strong>application backup policies</strong>, this creates a <strong>complete Kubernetes disaster recovery strategy</strong> for modern cloud-native environments.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3a36a360384c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Migrating Virtual Machines Across Namespaces in OpenShift Virtualization]]></title>
            <link>https://surajblog.medium.com/migrating-virtual-machines-across-namespaces-in-openshift-virtualization-497168518e1d?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/497168518e1d</guid>
            <category><![CDATA[migration]]></category>
            <category><![CDATA[openshift-virtualization]]></category>
            <category><![CDATA[vm-migration]]></category>
            <category><![CDATA[kubernetes-cluster]]></category>
            <category><![CDATA[red-hat-openshift]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Mon, 09 Mar 2026 17:18:41 GMT</pubDate>
            <atom:updated>2026-03-09T17:19:53.720Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sLICF0H8zHLW8iELo48R8w.png" /></figure><p>In enterprise OpenShift environments, namespaces typically represent lifecycle stages — <strong>development, pre-production, and production</strong>.</p><p>Promoting a Virtual Machine (VM) from one namespace to another often means cloning it in full — including its disks, networking configuration, and metadata — without disrupting consistency.</p><p>The <strong>Migration Toolkit for Virtualization (MTV)</strong>, built on the open-source <strong>Forklift project</strong>, makes this process seamless within the same OpenShift cluster.</p><p>In this guide, I’ll walk through a real-world, same-cluster namespace-to-namespace VM migration.</p><h3>What We’ll Cover</h3><p>Scope:</p><ul><li>Same cluster, namespace-to-namespace VM migration</li><li>Cold migration (full disk copy with VM stopped)</li><li>Handling shared bootable volumes across VMs</li></ul><p>Not covered:</p><ul><li>Cross-cluster migrations</li><li>VMware-to-OpenShift migrations</li><li>Warm (incremental) migrations</li></ul><h3>Prerequisites</h3><p>Before starting:</p><ul><li>OpenShift 4.14+</li><li>OpenShift Virtualization Operator installed</li><li>Migration Toolkit for Virtualization (MTV) Operator installed</li><li>Cluster admin access</li><li>Source namespace with VM(s)</li><li>Destination namespace created</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/566/1*xpEsxvtYKxfVP7ODedyMcA.png" /></figure><h3>Migration Workflow Overview</h3><p>Same-cluster migration is simpler than external migrations because:</p><ul><li>Both providers point to the same Kubernetes API</li><li>No external credentials required</li><li>Storage backend remains the same</li></ul><p>The process follows five core steps:</p><ol><li>Create Providers</li><li>Create Storage Map</li><li>Create Network Map</li><li>Create Migration Plan</li><li>Execute Migration</li></ol><h3>Step 1 — Create Providers</h3><p>Since this is a same-cluster migration, both providers point to the local Kubernetes API. No external credentials are needed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/746/1*jYFV7WAfNrmCjQIcKpz4fA.png" /></figure><pre>apiVersion: forklift.konveyor.io/v1beta1<br>kind: Provider<br>metadata:<br>  name: preprod<br>  namespace: openshift-mtv<br>spec:<br>  type: openshift<br>  secret: {}<br><br>---<br>apiVersion: forklift.konveyor.io/v1beta1<br>kind: Provider<br>metadata:<br>  name: prod<br>  namespace: openshift-mtv<br>spec:<br>  type: openshift<br>  secret: {}</pre><h3>Step 2 — Create Storage Map</h3><p>Maps storage classes between the source and the destination. Same cluster means same storage backend — values are identical.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/847/1*mIrpqZg5hlGiituxGC6gUA.png" /></figure><pre>apiVersion: forklift.konveyor.io/v1beta1<br>kind: StorageMap<br>metadata:<br>  name: storage-map<br>  namespace: openshift-mtv<br>spec:<br>  map:<br>    - source:<br>        storageClass: ocs-storagecluster-ceph-rbd-virtualization<br>      destination:<br>        storageClass: ocs-storagecluster-ceph-rbd-virtualization<br>  provider:<br>    source:<br>      name: source-provider<br>      namespace: openshift-mtv<br>    destination:<br>      name: destination-provider<br>      namespace: openshift-mtv</pre><h3>Step 3 — Create Network Map</h3><p>For pod networking:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/870/1*E7UHbaqA1kGOPndx3ZTKdA.png" /></figure><pre>apiVersion: forklift.konveyor.io/v1beta1<br>kind: NetworkMap<br>metadata:<br>  name: network-map<br>  namespace: openshift-mtv<br>spec:<br>  map:<br>    - source:<br>        type: pod<br>      destination:<br>        type: pod<br>  provider:<br>    source:<br>      name: source-provider<br>      namespace: openshift-mtv<br>    destination:<br>      name: destination-provider<br>      namespace: openshift-mtv<br></pre><p>For Multus bridge networking:</p><pre>spec:<br>  map:<br>    - source:<br>        name: default/localnet-network<br>        type: multus<br>      destination:<br>        name: default/localnet-network<br>        type: multus</pre><h3>Step 4 — Create the Migration Plan</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/924/1*pKxVB5eGDGGZ3F3aetBRbg.png" /></figure><pre>apiVersion: forklift.konveyor.io/v1beta1<br>kind: Plan<br>metadata:<br>  name: namespace-migration-plan<br>  namespace: openshift-mtv<br>spec:<br>  provider:<br>    source:<br>      name: source-provider<br>      namespace: openshift-mtv<br>    destination:<br>      name: destination-provider<br>      namespace: openshift-mtv<br>  map:<br>    network:<br>      name: network-map<br>      namespace: openshift-mtv<br>    storage:<br>      name: storage-map<br>      namespace: openshift-mtv<br>  targetNamespace: production<br>  warm: false<br>  runPreflightInspection: true<br>  preserveStaticIPs: false<br>  targetPowerState: auto<br>  useCompatibilityMode: true<br>  pvcNameTemplateUseGenerateName: true<br>  migrateSharedDisks: true<br>  vms:<br>    - name: my-application-vm<br>      namespace: pre-production</pre><h3>Step 5 — Execute the Migration</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/923/1*5sNO5G4irAaRyk203--JZQ.png" /></figure><pre>apiVersion: forklift.konveyor.io/v1beta1<br>kind: Migration<br>metadata:<br>  name: execute-migration<br>  namespace: openshift-mtv<br>spec:<br>  plan:<br>    name: namespace-migration-plan<br>    namespace: openshift-mtv</pre><h3>Monitoring Progress</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/743/1*eyg98G5Kb8kX-g6RVgtdog.png" /></figure><pre>oc get migration -n openshift-mtv -w<br>oc get plan -n openshift-mtv<br>oc get dv -n production<br>oc get vm -n production</pre><h3>Migrating Multiple VMs with a Shared Boot Volume</h3><p>When multiple VMs share a common bootable PVC (e.g., a Windows Server golden image), pre-stage it in the destination namespace before running the plan.</p><pre>apiVersion: cdi.kubevirt.io/v1beta1<br>kind: DataVolume<br>metadata:<br>  name: win-2k22<br>  namespace: production<br>spec:<br>  source:<br>    pvc:<br>      name: win-2k22<br>      namespace: pre-production<br>  storage:<br>    accessModes:<br>      - ReadWriteMany<br>    storageClassName: ocs-storagecluster-ceph-rbd-virtualization<br>    volumeMode: Block<br>    resources:<br>      requests:<br>        storage: 128Gi<br><br><br>oc get dv win-2k22 -n production -w<br># Wait for Phase -&gt; Succeeded</pre><p>Then include both VMs in the plan:</p><pre>  vms:<br>    - name: vm-1<br>      namespace: pre-production<br>    - name: vm-2<br>      namespace: pre-production</pre><p>MTV reuses the existing PVC instead of recreating it.</p><h3>Post-Migration Checklist</h3><pre>oc get vm -n production<br>virtctl start &lt;vm-name&gt; -n production<br>virtctl console &lt;vm-name&gt; -n production<br>oc get vmi -n production -o wide</pre><h3>Best Practices</h3><ul><li>Stop the source VM before migration to ensure data consistency.</li><li>Pre-clone shared volumes into the destination namespace before running the plan.</li><li>Use ReadWriteMany access mode for any PVC shared across multiple VMs.</li><li>Tag your plans with meaningful names for audit trails.</li><li>Monitor disk copy progress with oc get dv -n &lt;destination&gt; -w.</li></ul><h3>Conclusion</h3><p>MTV makes namespace-to-namespace VM migration on OpenShift straightforward — install the operator, define five resources in order, and your VMs land in the destination namespace with disks and networking intact. For same-cluster migrations, no credentials or external connectivity are needed.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=497168518e1d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kubernetes Descheduler Guide: Keep Your Cluster Balanced Automatically]]></title>
            <link>https://surajblog.medium.com/kubernetes-openshift-descheduler-guide-keep-your-cluster-balanced-automatically-17a703082eef?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/17a703082eef</guid>
            <category><![CDATA[descheduler]]></category>
            <category><![CDATA[kubernetes-cluster]]></category>
            <category><![CDATA[schedule]]></category>
            <category><![CDATA[openshift]]></category>
            <category><![CDATA[balance]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 16:50:22 GMT</pubDate>
            <atom:updated>2026-02-20T08:02:46.312Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Yz1WK5G5zt4z1zUMvgvXXQ.png" /></figure><h3>Introduction</h3><p>Have you ever noticed that your Kubernetes cluster becomes unbalanced over time? Some nodes end up running hundreds of pods while others sit nearly idle. This happens because the Kubernetes scheduler only makes decisions when pods are created — it doesn’t rebalance existing pods when cluster conditions change.</p><p>Enter the Descheduler — a tool that identifies pods that can be moved and evicts them so they can be rescheduled to better nodes.</p><p>In this article, I’ll share my hands-on experience configuring the Descheduler on OpenShift, including the challenges I faced and the solutions that actually worked.</p><h3>The Problem the Scheduler Doesn’t Solve</h3><p>The default scheduler only acts when pods are created. It does not rebalance running pods.</p><p>Common imbalance scenarios:</p><ul><li>Nodes recover after failure, but pods don’t return</li><li>New nodes are added, but remain underused</li><li>Resource usage drifts unevenly</li><li>Affinity or taint rules change</li><li>One node runs too many replicas</li></ul><p>Result: <strong>capacity waste + performance risk</strong></p><h3>What the Descheduler Does</h3><p>The Descheduler runs periodically and:</p><ol><li>Scans nodes and pods</li><li>Applies policy rules</li><li>Selects movable pods</li><li>Evicts them safely</li><li>Scheduler reschedules them</li></ol><p>Important: <strong>Descheduler does NOT place pods — it only evicts.</strong></p><h3>Useful Descheduler Strategies</h3><h3>Balance Strategies</h3><p>Used to spread pods better:</p><ul><li>LowNodeUtilization — move pods away from overloaded nodes</li><li>HighNodeUtilization — compact pods for scale-down</li><li>RemoveDuplicates — avoid the same ReplicaSet pods on one node</li><li>TopologySpread enforcement</li></ul><h3>Deschedule Strategies</h3><p>Used to remove invalid placements:</p><ul><li>Violating node affinity</li><li>Violating taints</li><li>Anti-affinity conflicts</li><li>Too many restarts</li><li>Very old pods</li><li>Failed pods</li></ul><h3>Example — Pod Count Balancing</h3><p>Cluster state: <br>Node1 — Pods 192<br>Node2 — Pods 92<br>Node3 — Pods 146</p><p>Goal: balance near ~140 pods each.</p><h3>Policy Example (Pod Count Based)</h3><pre>apiVersion: descheduler/v1alpha2<br>kind: DeschedulerPolicy<br><br>maxNoOfPodsToEvictPerNode: 20<br>maxNoOfPodsToEvictTotal: 50<br><br>profiles:<br>- name: BalancePods<br>  pluginConfig:<br>  - name: LowNodeUtilization<br>    args:<br>      thresholds:<br>        pods: 40   # &lt; 40% = underutilized (can receive pods)<br>      targetThresholds:<br>        pods: 68    # &gt; 68% = overutilized (evict from here)<br><br>  - name: DefaultEvictor<br>    args:<br>      nodeFit: false    # Allow eviction even if fit check fails<br>      minReplicas: 1<br>      evictLocalStoragePods: true<br><br>  plugins:<br>    balance:<br>      enabled:<br>      - LowNodeUtilization<br>    filter:<br>      enabled:<br>      - DefaultEvictor</pre><p>After several runs, the nodes balanced automatically.</p><h3>Key Lessons Learned</h3><h4>1. Don’t Mix Threshold Types</h4><p>Choose <strong>one</strong>:</p><pre># Option A: Pod count only<br>thresholds:<br>  pods: 40<br><br># Option B: Prometheus metrics only<br>thresholds:<br>  MetricResource: 0</pre><ul><li>Pod count</li><li>CPU/memory</li><li>Prometheus metrics</li></ul><p>Mixing causes policy errors.</p><h4>2. nodeFit Can Block Evictions</h4><p>If your pods have nodeSelector or specific tolerations, the descheduler may refuse to evict them because it can’t verify they’ll fit elsewhere.</p><p>Solution: Set nodeFit: false to let the scheduler handle placement.</p><pre>- name: DefaultEvictor<br>  args:<br>    nodeFit: false  # Trust the scheduler</pre><h4>3. Tune Eviction Limits for Your Environment</h4><pre>### Default limits are very conservative:<br><br>maxNoOfPodsToEvictPerNode: 2<br>maxNoOfPodsToEvictTotal: 5<br><br><br>#### For faster balancing:<br>maxNoOfPodsToEvictPerNode: 20<br>maxNoOfPodsToEvictTotal: 50</pre><h4>4. Pods That Are Protected</h4><p>Descheduler will not evict by default; these pods are protected:</p><ul><li>System-critical pods (kube-system)</li><li>DaemonSet pods</li><li>Pods with local storage</li><li>Pods violating PodDisruptionBudget</li><li>Static/mirror pods</li></ul><h3>Installation</h3><pre>helm repo add descheduler https://kubernetes-sigs.github.io/descheduler/<br>helm install descheduler descheduler/descheduler --namespace kube-system</pre><h3>Best Practices</h3><ul><li>Exclude critical namespaces</li><li>Test your policy before enabling actual evictions</li><li>Running as a CronJob is the recommended approach for production</li><li>Monitor eviction metrics</li><li>Set Appropriate Intervals</li><li>Monitor and Alert</li></ul><h3>Conclusion</h3><p>The descheduler keeps placement healthy over time. It’s not for constant movement; it’s for correcting drift and restoring balance. If you run medium-to-large clusters, the descheduler is not optional — it’s operational hygiene.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=17a703082eef" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[HAProxy Load Balancer Setup for OpenShift Bare-Metal Cluster (With VIP)]]></title>
            <link>https://surajblog.medium.com/haproxy-keepalived-load-balancer-for-openshift-bare-metal-with-vip-3c9ab32721a8?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/3c9ab32721a8</guid>
            <category><![CDATA[keepalived]]></category>
            <category><![CDATA[baremetal]]></category>
            <category><![CDATA[haproxy]]></category>
            <category><![CDATA[openshift]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Mon, 22 Dec 2025 15:25:37 GMT</pubDate>
            <atom:updated>2025-12-22T15:27:40.025Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*k1N-vyQRFQoxqRKrISlenw.png" /></figure><p>When running OpenShift on bare metal, we don’t have a managed load balancer, unlike those offered by AWS, Azure, or GCP. Instead, we typically deploy <strong>HAProxy + Keepalived</strong> to provide:</p><ul><li>High availability</li><li>Virtual IP (VIP) failover</li><li>Load balancing for control plane + ingress traffic</li></ul><p>In this blog, I’ll walk you through setting up HAProxy and Keepalived for an OpenShift bare-metal cluster.</p><h3>What We’ll Achieve</h3><ul><li>Two frontend load balancer nodes</li><li>A <strong>Virtual IP (VIP)</strong> floating between them</li><li><strong>Keepalived</strong> for failover</li><li><strong>HAProxy</strong> for routing OpenShift traffic</li><li>Full configuration examples</li></ul><h3>Prerequisites</h3><ul><li>OpenShift Bare-Metal cluster</li><li>Two Linux servers for Load Balancer</li><li>1 free IP for VIP</li></ul><h3>Step 1 — Install HAProxy &amp; Keepalived</h3><h3>On Both LB Nodes</h3><pre>sudo yum install haproxy keepalived -y</pre><p>Enable services:</p><pre>sudo systemctl enable haproxy<br>sudo systemctl enable keepalived</pre><h3>Step 2 — Configure HAProxy</h3><p>Open the HAProxy config:</p><pre>sudo vi /etc/haproxy/haproxy.cfg</pre><p>Paste the following:</p><pre>global<br>  log /dev/log local0<br>  log /dev/log local1 notice<br>  maxconn 20000<br>  daemon<br>defaults<br>  mode http<br>  log global<br>  retries 3<br>  timeout connect 10s<br>  timeout client 300s<br>  timeout server 300s<br><br># API Server Load Balance (6443)<br>frontend api<br>  bind *:6443<br>  default_backend api_backend<br>backend api_backend<br>  balance roundrobin<br>  server master01 192.168.1.201:6443 check<br>  server master02 192.168.1.202:6443 check<br>  server master03 192.168.1.203:6443 check<br># Machine Config Server (22623)<br>frontend mcs<br>  bind *:22623<br>  default_backend mcs_backend<br>backend mcs_backend<br>  balance roundrobin<br>  server master01 192.168.1.201:22623 check<br>  server master02 192.168.1.202:22623 check<br>  server master03 192.168.1.203:22623 check<br># Ingress HTTP (80)<br>frontend http_ingress<br>  bind *:80<br>  default_backend worker_http<br>backend worker_http<br>  balance roundrobin<br>  server worker01 192.168.1.211:80 check<br>  server worker02 192.168.1.212:80 check<br># Ingress HTTPS (443)<br>frontend https_ingress<br>  bind *:443<br>  default_backend worker_https<br>backend worker_https<br>  balance roundrobin<br>  server worker01 192.168.1.211:443 check<br>  server worker02 192.168.1.212:443 check<br><br><br># TCP direct forwarding from robot to openshift nodeport service<br><br>frontend robot_service<br>    bind *:30610<br>    mode tcp<br>    option tcplog<br>    default_backend robot_service<br><br>backend robot_service<br>    mode tcp<br>    balance source<br>    hash-type consistent<br>    server worker01 192.168.1.211:30610 check<br>    server worker02 192.168.1.212:30610 check</pre><p>Save and restart HAProxy:</p><pre>sudo systemctl restart haproxy<br>sudo systemctl status haproxy</pre><p>Note: If there is a requirement to pass traffic directly at the TCP layer, please configure the load balancer with mode tcp and balance source.</p><h3>Step 3 — Configure Keepalived (VIP Failover)</h3><h3>On LB Node 1 (MASTER)</h3><pre>sudo vi /etc/keepalived/keepalived.conf</pre><p>Paste the following:</p><pre>bal_defs {<br>  script_user root<br>  enable_script_security<br>  vrrp_version 3<br>  vrrp_min_garp true<br>}<br><br>vrrp_script chk_haproxy {<br>  script &quot;killall -0 haproxy&quot;<br>  fall 2<br>  rise 2<br>  interval 5<br>  weight 51<br>}<br><br>vrrp_instance VI_1 {<br>  state MASTER               # On secondary node use BACKUP<br>  interface bond0            # Or eth0 / ens192 depending on server<br>  virtual_router_id 29<br>  priority 100               # Lower (90) on secondary node<br>  advert_int 1<br><br>  virtual_ipaddress {<br>    10.21.154.16             # &lt;-- VIP<br>  }<br><br>  track_script {<br>    chk_haproxy<br>  }<br><br>  unicast_src_ip 10.21.154.11   # LB Node 1 IP<br>  unicast_peer {<br>    10.21.154.15                # LB Node 2 IP<br>  }<br>}</pre><h3>On LB Node 2 (BACKUP)</h3><pre>sudo vi /etc/keepalived/keepalived.conf</pre><pre>global_defs {<br>    process_names<br>    enable_script_security<br>    script_user root<br>    vrrp_version 3<br>    vrrp_min_garp true<br>}<br><br>vrrp_script chk_haproxy {<br>  script &quot;killall -0 haproxy&quot;<br>  fall 2<br>  rise 2<br>  interval 5<br>  weight 51<br>}<br>vrrp_instance VI_1 {<br>  interface bond0<br>  state BACKUP<br>  virtual_router_id 29<br>  priority 50<br>  virtual_ipaddress {<br>    10.21.154.16             # &lt;-- VIP<br><br>  }<br>  advert_int 5<br>  track_script {<br>    chk_haproxy<br>  }<br>  unicast_src_ip 10.21.154.15  # LB Node 2 IP<br>  unicast_peer {<br>     10.21.154.11    # LB Node 1 IP<br>  }<br>}</pre><h3>Start Keepalived</h3><p>On both nodes:</p><pre>sudo systemctl restart keepalived<br>sudo systemctl status keepalived</pre><h3>Verify VIP</h3><p>Run on master node:</p><pre>ip a | grep 10.21.154.16</pre><p>You should see VIP on MASTER.</p><p>Now test failover:</p><ul><li>Stop Keepalived on MASTER</li><li>VIP should move automatically to BACKUP</li></ul><pre>sudo systemctl stop keepalived</pre><p>Check again — VIP should now appear on LB2.</p><h3>Conclusion</h3><p>Setting up <strong>Keepalived with VRRP</strong> and <strong>HAProxy</strong> on an OpenShift bare-metal cluster provides a highly available and resilient front-end load-balancing solution. Keepalived ensures VIP failover between load balancer nodes, while HAProxy efficiently distributes traffic across the control plane and worker nodes. For HTTP/HTTPS workloads, HAProxy health checks help maintain reliability, and for applications that require direct TCP pass-through, using mode tcp with balance source ensures stable and session-consistent traffic handling.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3c9ab32721a8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My Journey to Cracking the AWS DevOps Professional Exam]]></title>
            <link>https://surajblog.medium.com/my-journey-to-cracking-the-aws-devops-professional-exam-fa72fe7452be?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/fa72fe7452be</guid>
            <category><![CDATA[preparation]]></category>
            <category><![CDATA[aws-devops]]></category>
            <category><![CDATA[exam-preparation]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[professional-development]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Thu, 28 Aug 2025 07:35:09 GMT</pubDate>
            <atom:updated>2025-08-28T10:18:51.518Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4ve56z9_QmhLTmBy-u0lRw.png" /></figure><p>I officially cleared the <strong>AWS Certified DevOps Engineer — Professional (DOP-C02)</strong> exam.<br> It wasn’t just about passing another certification — it was about <strong>proving to myself</strong> that my hands-on experience, persistence, and preparation could all come together for something meaningful.</p><p>I want to share my journey here because when I started, I was looking for stories like this to stay motivated. Hopefully, my path can inspire someone else who is preparing!</p><h3>Why I Took This Exam</h3><p>Working on multiple <strong>DevOps transformation projects</strong> at my company gave me confidence, but also showed me how broad the DevOps landscape is.</p><ul><li>I had the chance to work on <strong>project migrations</strong> and <strong>disaster recovery (DR) setups</strong>, which taught me resilience and problem-solving under pressure.</li><li>Handling <strong>multi-account deployments</strong> and implementing <strong>AWS Control Tower</strong> gave me exposure to advanced-level AWS organisation and governance.</li><li>Designing <strong>deployment processes</strong> across environments (Dev, QA, Prod) taught me the importance of automation, repeatability, and security.</li></ul><p>I realised the <strong>AWS DevOps Professional certification</strong> aligned perfectly with the kind of work I was already doing — so I decided to challenge myself.</p><h3>Preparation Strategy</h3><p>I broke my prep into three phases:</p><h3>1. Learning the Core Concepts</h3><p>I enrolled in <a href="https://www.udemy.com/course/aws-certified-devops-engineer-professional-hands-on/?utm_source=chatgpt.com"><strong>Stephane Maarek’s AWS Certified DevOps Engineer — Professional (Hands-on)</strong></a> on Udemy.</p><p>Stephane’s course was my <strong>foundation</strong>:</p><ul><li>Crisp explanations of complex concepts</li><li>Hands-on labs that reinforced real-world scenarios</li><li>Coverage of exam domains in a very structured way</li></ul><p>Whenever I felt I needed more depth, I jumped into <strong>AWS documentation</strong> — especially for services like <strong>CloudFormation, CodePipeline, OpsWorks, and Elastic Beanstalk</strong>.</p><h3>2. Practice Tests &amp; Self-Assessment</h3><p>Once I built a strong foundation, I shifted focus to practice.</p><p>I purchased <a href="https://www.udemy.com/course/aws-certified-devops-engineer-professional-practice-exams-amazon-dop-c02?utm_source=chatgpt.com"><strong>Jon Bonso’s Practice Exams (Tutorials Dojo)</strong></a>. And let me tell you — <strong>these were a game changer</strong>.</p><p>Here’s how I used them:</p><ul><li>Took a full-length test, reviewed every wrong answer.</li><li>Didn’t stop at “what’s correct” but focused on <strong>why other options were wrong</strong>.</li><li>Built a habit of mapping scenarios to <strong>real-world use cases</strong> I’d already worked on.</li></ul><p>This not only improved my scores but also boosted my <strong>confidence for scenario-based questions</strong> in the exam.</p><h3>3. Revision &amp; Real-Work Mapping</h3><p>In the last few weeks, I made daily revision lists for:</p><ul><li>CI/CD pipelines and automation patterns</li><li>IAM best practices &amp; cross-account access</li><li>Monitoring &amp; logging (CloudWatch, X-Ray)</li><li>Deployment strategies in <strong>Elastic Beanstalk</strong> and <strong>ECS/EKS</strong></li><li>Governance with <strong>Control Tower &amp; Organisations</strong></li><li><strong>Automation setup</strong> (Lambda, Step Functions, EventBridge rules)</li><li><strong>Infrastructure as Code</strong> with CloudFormation</li><li><strong>Notification systems</strong> (SNS, Slack integration, custom alerts)</li><li><strong>Auto remediation</strong> workflows triggered by security events or CloudWatch alarms</li></ul><p>At the same time, I connected topics back to <strong>my project work</strong>. For example:</p><ul><li>While revising <strong>multi-account CI/CD</strong>, I recalled how I set up GitHub Actions with AWS CodePipeline for deployments across multiple environments.</li><li>When studying <strong>disaster recovery (DR) strategies</strong>, I connected it to actual DR drills I had performed at work.</li><li>While learning <strong>EventBridge event-driven automation</strong>, I mapped it to real-world event handling → automatically restarting failed ECS tasks or notifying teams via Slack.</li><li>On <strong>EC2 Auto Scaling Groups</strong>, I practised setting up <strong>automated scaling with multiple deployment strategies (Blue/Green, Canary, Rolling)</strong> to ensure <strong>zero downtime</strong> and even tested <strong>rollback automation</strong> when health checks failed.</li><li>With <strong>CloudFormation IAC</strong>, I recalled implementing repeatable environment setups for DR projects.</li></ul><p>That <strong>real-work connection</strong> made all the difference.</p><h3>Exam Day Experience</h3><p>I took the exam from home via <strong>Pearson VUE</strong>.<br> Big shoutout to them — it was a seamless experience, no technical hiccups, and very professional proctoring.</p><p>During the exam:</p><ul><li>I flagged tricky questions but trusted my first instinct in most cases.</li><li>Time management was key — I finished with around 15 minutes to review flagged questions.</li><li>The exam is tough, but if you’ve <strong>worked in real DevOps projects</strong> and practised well, it feels like applying logic rather than guessing.</li></ul><h3>Key Takeaways</h3><ul><li><strong>Practical Experience &gt; Theory</strong>: My hands-on work with migrations, DR, and deployments made concepts click faster.</li><li><strong>Practice Tests are GOLD</strong>: Jon Bonso’s tests sharpened my exam mindset.</li><li><strong>Stay Consistent</strong>: Even 1–2 hours daily over weeks matters more than last-minute cramming.</li><li><strong>Map to Real Work</strong>: Always connect AWS services to actual use cases.</li></ul><h3>Final Words</h3><p>Passing <strong>AWS DevOps Engineer — Professional</strong> is not just about memorising services. It’s about:</p><ul><li><strong>Designing with resilience</strong></li><li><strong>Automating everything</strong></li><li><strong>Thinking at scale across accounts and environments</strong></li></ul><p>If you’re preparing, my advice is: <strong>trust the process, combine theory with practice, and believe in your hands-on skills.</strong></p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fa72fe7452be" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Installing DevLake for KPI Management to Track GitHub PRs, Jira Efforts and More]]></title>
            <link>https://surajblog.medium.com/installing-devlake-for-kpi-management-to-track-github-prs-jira-efforts-and-more-e82259fb29d4?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/e82259fb29d4</guid>
            <category><![CDATA[kpi-tracking]]></category>
            <category><![CDATA[installation]]></category>
            <category><![CDATA[kpi-management]]></category>
            <category><![CDATA[apache-devlake]]></category>
            <category><![CDATA[kpi]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Mon, 21 Jul 2025 18:07:24 GMT</pubDate>
            <atom:updated>2025-07-23T05:47:43.560Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L0Bl5tgZgowXcBA0Kal56A.png" /></figure><p><a href="https://devlake.apache.org/">DevLake (Dev Data Lake)</a> is an open-source platform built on Apache, helping engineering teams measure software delivery performance using data from GitHub, Jira, GitLab, Jenkins, and other sources.</p><p>If you want to track DORA metrics, PR velocity, lead time, cycle time, or custom engineering KPIs — DevLake gives you powerful dashboards out of the box.</p><h3>Why Choose DevLake?</h3><ul><li>Unified Data Aggregation: Connects tools like GitHub and Jira for comprehensive analysis.</li><li>Out-of-the-Box Dashboards: Prebuilt charts for key metrics such as Pull Request (PR) Pickup Time, Review Time, Issue Throughput, and more.</li><li>Flexible &amp; Customizable: Extendable with new data sources, metrics, and SQL-based dashboards.</li><li>Supports DORA &amp; Agile Metrics: Instant metrics for PRs, code review, and agile planning.</li></ul><h3>Prerequisites</h3><p>We’ll install DevLake on a <strong>single server</strong> with Docker Compose. Here’s what you need:</p><ul><li>A Linux-based host (tested on Ubuntu/RHEL/Amazon Linux)</li><li>At least <strong><em>2 vCPUs, 4 GB RAM</em></strong>, and <strong><em>10GB</em></strong> disk space</li><li><strong><em>Docker</em></strong><em> </em>and <strong><em>Docker-Compose</em></strong> installed</li></ul><p>For integrations:</p><ul><li>GitHub: A Personal Access Token (PAT) and Repo URL</li><li>Jira: API token, Jira domain, and user email</li></ul><h3>Step-by-Step Installation Guide</h3><h3>1. Download Latest Stable Release</h3><p>Visit: <a href="https://github.com/apache/incubator-devlake/releases">https://github.com/apache/incubator-devlake/releases</a></p><p>Replace version in the URLs below with the <strong>latest stable tag</strong>.</p><pre>mkdir devlake &amp;&amp; cd devlake<br><br># Download latest stable docker-compose and env config<br>wget https://github.com/apache/incubator-devlake/releases/download/v1.0.2/docker-compose.yml<br>wget https://github.com/apache/incubator-devlake/releases/download/v1.0.2/env.example</pre><h3>2. Generate ENCRYPTION_SECRET</h3><pre>## This key encrypts tokens and passwords in your DB. Keep it safe!<br><br>openssl rand -base64 2000 | tr -dc &#39;A-Z&#39; | fold -w 128 | head -n 1</pre><p>Copy the 128-character uppercase string and paste it into .env:</p><pre>ENCRYPTION_SECRET=YOUR_SECURE_STRING</pre><h3>3. Rename and Edit .env File</h3><pre>mv env.example .env</pre><p>Now open .env and scroll to:</p><pre># Sensitive information encryption key<br>ENCRYPTION_SECRET=</pre><h3>4. Launch DevLake</h3><p>Start the stack:</p><pre>docker-compose up -d</pre><p>You’ll see Docker spinning up the following services:</p><pre>&gt; devlake (core backend)<br>&gt; mysql (database)<br>&gt; grafana (dashboards)<br>&gt; config-ui (to configure plugins and connections)</pre><p>Check if services are running:</p><pre>docker ps</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zQxDkbTYrgSGAQMkXxJ2iQ.png" /></figure><h3>5. Verify DevLake UI:</h3><ul><li>Open your browser and navigate to Config UI: <a href="http://public_ip:4000/">http://public_ip:4000</a></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nU8MMlQLsuh5eG9-s3HEZg.png" /></figure><p><strong>Note:</strong> If config-ui says “API unreachable,” wait 10–15 seconds and refresh.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NCmu4kJikInMRWlzjnvaeg.png" /></figure><h3>Step 6: Setup and Configuration</h3><h4><strong>Onboarding:</strong></h4><p>When you access the Config UI, you’ll be prompted to create your first project.</p><ul><li>Name your project and define its scope — such as “GitHub + Jira Metrics”.</li><li>Security &amp; Authentication (Optional):</li></ul><p>DevLake supports environment variable–based authentication, which you can adjust in .env config as required.</p><h4>Connect GitHub</h4><ul><li>Go to Data Connections → Add Connection → Select GitHub.</li><li>Enter Name: Any identifier.</li><li>API URL: <a href="https://api.github.com/">https://api.github.com/</a></li><li><strong>If you are using GitHub Cloud, you do not need to enter the endpoint URL, which is </strong><a href="https://api.github.com/"><strong>https://api.github.com</strong></a></li><li>Personal Access Token: Create from your GitHub account settings (repo, read:user scope is sufficient.).</li><li>Test and Save the connection.</li></ul><h4>Connect Jira</h4><ul><li>Go to Data Connections → Add Connection → Select Jira.</li><li>Enter: Name: Any identifier.</li><li>Endpoint URL: Your Jira domain (https://your-company.atlassian.net/.)</li><li>Email: Your Jira login</li><li>API Token: Obtain from Atlassian account settings.</li><li>Test and Save.</li></ul><h3>Step 7: Creating a Project and Blueprint</h3><p><strong>Create a Blueprint:</strong></p><ul><li>Define which repos and Jira projects/issues you want to sync.</li><li>By default, DevLake collects data for the last 14 days; adjust this setting for broader data collection if needed.</li></ul><p><strong>Configure Data Scope:</strong></p><ul><li>For GitHub: Select repositories, and types of data (PRs, Issues, etc.).</li><li>For Jira: Select projects, issue types, or other relevant entities.</li></ul><p><strong>Set Up Scheduling:</strong></p><ul><li>Configure sync to run at preferred intervals, or trigger manually for one-time pulls.</li></ul><p><strong>Start Data Collection:</strong></p><ul><li>Run the blueprint to begin aggregation.</li></ul><h3>Step 8: Exploring KPI Dashboards</h3><ul><li>Grafana Dashboards: Access prebuilt dashboards at <a href="http://localhost:3002/">http://localhost:3002</a> (Default: admin/admin).</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JFATUPEgrsmxskZYdDafxg.png" /></figure><h3>Track Engineering KPIs</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lQtNN7hBoqpKsQZGePs89Q.png" /></figure><p>Key PR KPIs:</p><ul><li>PR Pickup Time: Measures speed from PR creation to first review.</li><li>PR Review Time: From first response until PR merged.</li><li>PR Volume &amp; Merge Rates: Track contributions by developer, team, and interval.</li></ul><p>Jira KPI Highlights:</p><ul><li>Issue Throughput: Number of issues completed per interval.</li><li>Story Point Tracking: Effort estimation vs. actual velocity.</li></ul><p>Customise Dashboards:</p><ul><li>Modify or add new panels in Grafana.</li><li>Run SQL queries to create custom insights.</li></ul><h3>Conclusion</h3><p>With Apache DevLake, you can aggregate, analyse, and visualise KPIs spanning Github, Jira, and other development tools. In minutes, you’ll have dashboards that give clarity on PR activity, code review health, Jira velocity, and more.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer 👏</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e82259fb29d4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Attach a Secondary Network Interface to VMs in OpenShift without a bond]]></title>
            <link>https://surajblog.medium.com/how-to-attach-a-secondary-network-interface-to-vms-in-openshift-d28c0c72f402?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/d28c0c72f402</guid>
            <category><![CDATA[netwroking]]></category>
            <category><![CDATA[multus]]></category>
            <category><![CDATA[openshift]]></category>
            <category><![CDATA[network-operator]]></category>
            <category><![CDATA[openshift-virtualization]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Thu, 17 Jul 2025 07:22:59 GMT</pubDate>
            <atom:updated>2025-08-15T08:38:46.798Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rcbdATCjYzEcKdywKh-dQQ.png" /></figure><p>In some OpenShift environments, certain virtual machines (VMs) need to bypass the default Pod network and connect directly to the node’s physical network (L2 segment) — for example, to access external switches, legacy systems, or devices on the factory floor.</p><p>The default OpenShift pod network (OVN-Kubernetes) performs masquerading (NAT) and may block or rewrite traffic, making it unsuitable for direct external communication.</p><p>The solution: create a Linux bridge over a physical NIC, define a Multus network over it, and assign static IP addresses from the physical network to your VM via the secondary interface.</p><h3>Prerequisites</h3><ul><li>OpenShift cluster with OpenShift Virtualisation and Multus enabled</li><li>A physical NIC not used by cluster networking (e.g. eno12329np1)</li><li>Administrative access to apply NodeNetworkConfigurationPolicy and NAD</li></ul><h3>Step 1: Create a Linux Bridge Over the Physical NIC</h3><p>We will use <strong>NMState</strong> to create a br0 bridge on a spare NIC (e.g., eno12409np1).</p><pre>oc apply -f nncp-bridge.yaml<br><br>apiVersion: nmstate.io/v1<br>kind: NodeNetworkConfigurationPolicy<br>metadata:<br>  name: eno12329np1-nncp<br>spec:<br>  desiredState:<br>    interfaces:<br>      - name: br0<br>        type: linux-bridge<br>        state: up<br>        bridge:<br>          port:<br>            - name: eno12329np1 ( physical interface)<br>          options:<br>            stp:<br>              enabled: false<br>  nodeSelector:<br>    nmstate: &quot;true&quot;</pre><p>Wait for the node to apply the bridge:</p><pre>oc get mcp</pre><p><strong>Note: </strong>This will not work when a bond is configured. We need to follow another configuration for NNCP</p><h3>Step 2: Define a NetworkAttachmentDefinition (NAD)</h3><p>This configures a Multus network named nad, which uses the br0 bridge and assigns a <strong>static IP</strong>.</p><pre>oc apply -f nad.yaml<br><br>apiVersion: k8s.cni.cncf.io/v1<br>kind: NetworkAttachmentDefinition<br>metadata:<br>  name: nad<br>  namespace: my-namespace # where you want to attach it<br>spec:<br>  config: |-<br>    {<br>      &quot;cniVersion&quot;: &quot;0.3.1&quot;,<br>      &quot;name&quot;: &quot;nad&quot;,<br>      &quot;type&quot;: &quot;bridge&quot;,<br>      &quot;bridge&quot;: &quot;br0&quot;,<br>      &quot;ipam&quot;: {<br>        &quot;type&quot;: &quot;static&quot;,<br>        &quot;addresses&quot;: [<br>          {<br>            &quot;address&quot;: &quot;10.41.115.14/24&quot;, #Replace the IP address and gateway with your physical VLAN/subnet settings.<br>            &quot;gateway&quot;: &quot;10.41.115.1&quot;<br>          }<br>        ]<br>      }<br>    }</pre><p>Replace the IP address and gateway with your physical VLAN/subnet settings.</p><h3>Step 3: Attach This Secondary Interface to a VM</h3><p>In your VirtualMachine YAML, add the Multus network under networks, and map it under interfaces.</p><pre>spec:<br>  networks:<br>    - name: nad<br>      multus:<br>        networkName: my-vm-namespace/nad<br>domain:<br>    devices:<br>      interfaces:<br>        - name: nad<br>          bridge: {}</pre><p>You can configure the interface inside the VM manually or use cloud-init to assign a static IP.</p><h3>Step 4: Verify</h3><p>Once the VM is running:</p><pre>oc rsh vm-name<br>ip a</pre><p>You should see the secondary interface (e.g., eth1) with the static IP 10.41.125.14.</p><h3>Use Case</h3><ul><li>Direct communication with external physical devices and factory equipment.</li><li>Avoid pod network NAT or firewall restrictions</li><li>Isolate workloads using VLANs or external routing</li></ul><h3>Conclusion</h3><p>When your OpenShift workloads or VMs need direct access to external Layer 2 networks, the default pod network falls short due to NAT and security rules. By setting up a secondary interface using a Linux bridge over a dedicated NIC, you can cleanly separate external and internal traffic, enable direct VM communication on your data centre VLAN, and avoid dependency on bonding or additional NICs.</p><p>If you have any questions or feedback, please don&#39;t hesitate to comment.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer 👏</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d28c0c72f402" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Installing Veeam Kasten K10 on OpenShift Bare Metal with OAuth Integration]]></title>
            <link>https://surajblog.medium.com/installing-veeam-kasten-k10-on-openshift-bare-metal-with-openshift-oauth-integration-afffc527fe86?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/afffc527fe86</guid>
            <category><![CDATA[s3-bucket]]></category>
            <category><![CDATA[oauth]]></category>
            <category><![CDATA[backup-and-restore]]></category>
            <category><![CDATA[veeam-backup-software]]></category>
            <category><![CDATA[openshift]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Mon, 19 May 2025 17:41:11 GMT</pubDate>
            <atom:updated>2026-03-16T11:16:30.972Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t0XuqsPWCbeWY0IIrHhFVA.png" /></figure><h3>Overview</h3><p>In this guide, I walk through a full installation of <a href="https://www.kasten.io/">Veeam Kasten K10</a> on an OpenShift bare-metal cluster with OpenShift OAuth integration, TLS certificate setup, and custom CA bundle configuration. This setup ensures enterprise-grade resiliency and compliance if you’re managing data protection and backup on a Kubernetes/OpenShift environment.</p><h3>Step 1: Install Kasten K10 Operator</h3><p>Head to the <strong>OperatorHub</strong> inside OpenShift and search for:</p><p><em>Veeam Kasten (Enterprise - Term)</em></p><p>Install the operator in a dedicated namespace:</p><pre>oc new-project kasten-io</pre><p>From the UI, set the <strong>installation namespace</strong> as kasten-io.</p><h3>Step 2: Create OpenShift OAuth-Compatible Service Account</h3><p>Create a service account annotated with a redirect URI for OAuth login via the OpenShift console:</p><pre>apiVersion: v1<br>kind: ServiceAccount<br>metadata:<br>  name: k10-dex-sa<br>  namespace: kasten-io<br>  annotations:<br>    serviceaccounts.openshift.io/oauth-redirecturi.dex: https://k10-route-kasten-io.apps.&lt;your-domain&gt;/k10/dex/callback</pre><p>Apply it:</p><pre>oc apply -f k10-serviceaccount.yaml</pre><h3>Step 3: Create Token Secret for OAuth</h3><p>This secret is required to extract the token for integration with OpenShift identity providers:</p><pre>apiVersion: v1<br>kind: Secret<br>type: kubernetes.io/service-account-token<br>metadata:<br>  name: k10-dex-sa-secret<br>  namespace: kasten-io<br>  annotations:<br>    kubernetes.io/service-account.name: &quot;k10-dex-sa&quot;</pre><p>Apply it:</p><pre>oc apply -f k10-secret.yaml</pre><h3>Step 4: Retrieve the Token from the Secret</h3><p>Now, extract the actual token:</p><pre>kubectl -n kasten-io get secret k10-dex-sa-secret -o jsonpath=&#39;{.data.token}&#39; | base64 -d &amp;&amp; echo</pre><p>Copy this token; it will be used in the K10 manifest.</p><h3>Step 5: Handle OpenShift Cluster TLS Certificates</h3><p>Export the OpenShift <strong>router CA</strong>:</p><pre>oc get secret router-ca -n openshift-ingress-operator -o jsonpath=&#39;{.data.tls\.crt}&#39; | base64 --decode &gt; router-ca.pem</pre><p>Export the <strong>console certificate</strong>:</p><pre>openssl s_client -showcerts -connect console-openshift-console.apps.&lt;your-domain&gt;:443 &lt;/dev/null 2&gt;/dev/null | \<br>openssl x509 -outform PEM &gt; openshift-console-cert.pem</pre><p>Combine both:</p><pre>cat openshift-console-cert.pem router-ca.pem &gt; custom-ca-bundle.pem</pre><p>Create a ConfigMap:</p><pre>oc -n kasten-io create configmap custom-ca-bundle-store --from-file=custom-ca-bundle.pem</pre><h3>Step 6: Deploy Kasten K10 Instance</h3><p><strong>Prerequistes</strong></p><pre>$ kubectl annotate volumesnapshotclass \ <br>ocs-storagecluster-cephfsplugin-snapclass \<br> k10.kasten.io/is-snapshot-class=true<br> <br>$ kubectl annotate volumesnapshotclass \<br>ocs-storagecluster-rbdplugin-snapclass \<br>   k10.kasten.io/is-snapshot-class=true</pre><p>Replace the placeholder token and domain values accordingly in the below YAML:</p><pre>apiVersion: apik10.kasten.io/v1alpha1<br>kind: K10<br>metadata:<br>  name: k10<br>  namespace: kasten-io<br>  annotations:<br>    helm.sdk.operatorframework.io/reconcile-period: 2m<br>    helm.sdk.operatorframework.io/rollback-force: &#39;false&#39;<br>  finalizers:<br>    - helm.sdk.operatorframework.io/uninstall-release<br>spec:<br>  auth:<br>    openshift:<br>      enabled: true<br>      serviceAccount: &quot;k10-dex-sa&quot;<br>      clientSecret: &quot;&lt;your-oauth-token&gt;&quot;<br>      dashboardURL: &quot;https://k10-route-kasten-io.apps.&lt;your-domain&gt;/k10&quot;<br>      openshiftURL: &quot;https://api.&lt;your-domain&gt;:6443&quot;<br>      insecureCA: false<br>  cacertconfigmap:<br>    name: custom-ca-bundle-store<br>  global:<br>    persistence:<br>      storageClass: &quot;ocs-storagecluster-ceph-rbd&quot;<br>  route:<br>    enabled: true<br>    tls:<br>      enabled: true</pre><p>Apply:</p><pre>oc apply -f k10-instance.yaml</pre><pre>$ kubectl annotate volumesnapshotclass<br>ocs-storagecluster-cephfsplugin-snapclass \<br>   k10.kasten.io/is-snapshot-class=true<br> <br>$ kubectl annotate volumesnapshotclass<br>ocs-storagecluster-rbdplugin-snapclass \<br>   k10.kasten.io/is-snapshot-class=true</pre><h3>Validation</h3><ul><li>Visit the dashboard: https://k10-route-kasten-io.apps.&lt;your-domain&gt;/k10</li><li>You should be redirected to the OpenShift login screen.</li><li>Once authenticated, you’ll land inside the <strong>Kasten K10 UI</strong>.</li></ul><h3>Security Note</h3><ul><li>Make sure you restrict token permissions to minimal roles.</li><li>Secure the service account and limit user impersonation.</li><li>Keep your CA bundle updated if OpenShift ingress changes.</li></ul><h3>Metrics exports:</h3><p>If you want to export metrics to Datadog or an external system, please use the spec below spec.prometheus config.</p><pre>apiVersion: apik10.kasten.io/v1alpha1<br>kind: K10<br>metadata:<br>  name: k10<br>  namespace: kasten-io<br>spec:<br>  prometheus:<br>    server:<br>      podAnnotations:<br>        ad.datadoghq.com/prometheus-server.check_names: |<br>          [&quot;openmetrics&quot;]<br>        ad.datadoghq.com/prometheus-server.init_configs: |<br>          [{}]<br>        ad.datadoghq.com/prometheus-server.instances: |-<br>          [{<br>            &quot;prometheus_url&quot;: &quot;http://%%host%%:%%port%%/k10/prometheus/federate?match[]=%7Bjob%3D~%22.%2B%22%7D&quot;,<br>            &quot;namespace&quot;: &quot;kasten-io&quot;,<br>            &quot;metrics&quot;: [<br>              &quot;action_export_*&quot;,<br>              &quot;action_backup_*&quot;     <br>            ],<br>            &quot;type_overrides&quot;: {&quot;*&quot;: &quot;counter&quot;}<br>          }]<br></pre><h3>Conclusion</h3><p>With this setup, Veeam Kasten K10 is now successfully running on your OpenShift bare metal cluster, integrated with OpenShift OAuth for secure access and custom CA certificates for TLS validation.</p><p>This forms the foundation for managing data protection, backup, and disaster recovery in your Kubernetes workloads.</p><p><strong>Next up</strong>: In the upcoming blog, we’ll walk you through how to <strong>add target backup storage,</strong> such as NFS, S3-compatible object storage, or enterprise storage backends, to complete the end-to-end backup solution using Kasten K10.</p><p>If you have any questions or feedback, please don&#39;t hesitate to comment.</p><p>:) Be sure to <strong>clap</strong> and <strong>follow</strong> the writer 👏</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=afffc527fe86" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Exposing Fixed Ports to External Applications in OpenShift Bare Metal Using External IP’S]]></title>
            <link>https://surajblog.medium.com/exposing-fixed-ports-to-external-applications-in-openshift-bare-metal-using-external-ips-7d8031c467e1?source=rss-b6f5f0f84810------2</link>
            <guid isPermaLink="false">https://medium.com/p/7d8031c467e1</guid>
            <category><![CDATA[load-balancer]]></category>
            <category><![CDATA[openshift]]></category>
            <category><![CDATA[networking]]></category>
            <category><![CDATA[external-ip]]></category>
            <category><![CDATA[robotics-automation]]></category>
            <dc:creator><![CDATA[Suraj Solanki]]></dc:creator>
            <pubDate>Fri, 09 May 2025 10:30:31 GMT</pubDate>
            <atom:updated>2025-08-11T14:40:09.662Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qfO9sjUFrIcWsodVxvN00g.png" /></figure><p>In a recent challenge we faced while operating our applications in an OpenShift bare metal environment, we had to figure out how to expose internal services to a robot application running <strong>outside the cluster</strong>. This presented constraints that made typical exposure models, such as NodePort or LoadBalancer — <strong>ineffective</strong></p><h3>The Problem</h3><p>We had an application running inside our OpenShift cluster that needed to communicate with a robot client outside the cluster. The robot was configured to connect to a fixed port range (e.g., 5000–5008), and modifying these ports wasn’t an option.</p><p>However NodePort services only expose ports in a restricted range (default: 30000–32767).</p><p>To complicate things further:</p><ul><li>The robot was hardcoded to use specific ports outside this range.</li><li>We had to ensure <strong>direct accessibility</strong> from the external client without modifying port mappings.</li><li>LoadBalancer services weren’t viable in our bare-metal setup without a cloud provider integration.</li></ul><p>So we needed an alternative approach.</p><h3>The Solution: External IPs</h3><p>We solved the problem using <strong>External IPs</strong>, which allow Kubernetes services to be exposed directly via <strong>node IP addresses</strong> and <strong>arbitrary ports</strong> — even outside the default NodePort range.</p><h3>What We Did</h3><h4>1. Identify a Reachable Node IP</h4><p>We selected a worker node with a reachable IP (one that the robot network could connect to):</p><pre>oc get nodes -o wide<br><br>worker-node1   Ready    worker   ...   INTERNAL-IP: 192.168.100.101</pre><h4>2. Allow the Node IP in the External IP Policy</h4><p>OpenShift restricts the use of external IPs by default. To allow our node IP, we edited the cluster network configuration:</p><pre>oc edit network.config cluster<br><br>spec:<br>  externalIP:<br>    policy:<br>      allowedCIDRs:<br>        - 192.168.100.101/32   # Allow exact node IP</pre><p>Under the spec.externalIP.policy section:</p><p><em>This configuration tells OpenShift to </em><strong><em>allow only the specified IP</em></strong><em> to be used for external access.</em></p><h4>3. Create a Service with External IP</h4><p>We defined a ClusterIP service with an externalIPs field:</p><p>Verify :</p><pre>ssh user@nodeip<br><br>check the listening port <br><br>$ss -tunl | grep -i 5000</pre><pre>oc apply -f robot-connector-service.yaml<br><br>apiVersion: v1<br>kind: Service<br>metadata:<br>  name: robot-service-external<br>  namespace: my-app<br>spec:<br>  selector:<br>    app: my-application-vm<br>  type: ClusterIP<br>  ports:<br>    - protocol: TCP<br>      port: 5000         # Pod port<br>      targetPort: 5000   # Robot expects this port<br>  externalIPs:<br>    - 192.168.100.101   # Node IP allowed in the policy<br></pre><h4>4. Robot Configuration</h4><p>The robot was configured to connect directly to the <strong>Node IP</strong> on its fixed ports:</p><pre>curl http://&lt;NodeIP&gt;:5000</pre><p>The traffic reached theexternalIP, which OpenShift routed to the right pod.</p><p><strong><em>Note:</em></strong><em> Make sure port 5000 is not already used on the node. If another service (e.g., a system daemon or container runtime) is already bound to that port, the traffic will not be forwarded correctly and may cause connection failures.</em></p><h3>Why This Worked</h3><ul><li><strong>External IPs</strong> allow Kubernetes to listen on <strong>any port</strong> on the node — not limited to the NodePort range.</li><li>The robot app didn’t need modification.</li><li>No need for a cloud-based LoadBalancer or Ingress Controller.</li></ul><h3>Considerations</h3><ul><li>You need to ensure <strong>firewall rules</strong> allow traffic on those ports on the node.</li><li>Services using external IPs are <strong>not load balanced</strong> unless you configure it at a lower network layer (e.g., keepalived, HAProxy).</li><li>Node-level availability is important. If the selected node goes down, so does external access to that port.</li></ul><h3>Conclusion</h3><p>This use case illustrates how <strong>External IPs in OpenShift</strong> can be a powerful tool when dealing with constraints that fall outside typical Kubernetes service patterns.</p><p>While not always ideal for large-scale scenarios, they provide a pragmatic and effective way to expose specific ports, especially when legacy or external systems are involved and configuration changes are limited.</p><p>If you have any questions or feedback, feel free to comment.</p><p><strong>About The Author</strong><br>Suraj Solanki<br>Senior DevOps Engineer<br>LinkedIn: <a href="https://www.linkedin.com/in/suraj-solanki/">https://www.linkedin.com/in/suraj-solanki</a><br>Topmate: <a href="https://topmate.io/suraj_solanki">https://topmate.io/suraj_solanki</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d8031c467e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>