<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Vitalii Shumylo</title>
    <description>The latest articles on DEV Community by Vitalii Shumylo (@vitalii_shumylo).</description>
    <link>https://dev.to/vitalii_shumylo</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4200160%2Fdc588a61-73d8-4983-aee5-6daf7eeb1444.png</url>
      <title>DEV Community: Vitalii Shumylo</title>
      <link>https://dev.to/vitalii_shumylo</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/vitalii_shumylo"/>
    <language>en</language>
    <item>
      <title>The Online Responder says OK, but certutil shows no OCSP check: look at the certificate</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Fri, 13 Nov 2026 10:03:36 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/the-online-responder-says-ok-but-certutil-shows-no-ocsp-check-look-at-the-certificate-5og</link>
      <guid>https://dev.to/vitalii_shumylo/the-online-responder-says-ok-but-certutil-shows-no-ocsp-check-look-at-the-certificate-5og</guid>
      <description>&lt;p&gt;You set up an Online Responder on Windows Server. The console shows a green status. But when you test a certificate with &lt;code&gt;certutil&lt;/code&gt;, the output lists only CRL checks. There is no OCSP line, and clients still download the full CRL.&lt;/p&gt;

&lt;p&gt;The cause is usually the certificate, not the responder. A client cannot search for a responder. It only uses the address that is written inside the certificate. The CA adds that address when it issues the certificate. If you change the CA's AIA settings later, certificates that already exist do not change. So a test certificate from last week has no OCSP address, and the client goes back to the CRL. Also check the option on the AIA entry: the address must go into the OCSP extension, not the normal AIA extension.&lt;/p&gt;

&lt;p&gt;Add the address on the CA, then test with a certificate issued after the change:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight powershell"&gt;&lt;code&gt;&lt;span class="c1"&gt;# On the CA: write the responder address into the OCSP extension of new certificates&lt;/span&gt;
&lt;span class="nc"&gt;Add&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="nc"&gt;CAAuthorityInformationAccess&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="nc"&gt;AddToCertificateOcsp&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="nc"&gt;Uri&lt;/span&gt; &lt;span class="s1"&gt;'http://ocsp.corp.learnitlessons.com/ocsp'&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="nc"&gt;Force&lt;/span&gt;
&lt;span class="nc"&gt;Restart&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="nc"&gt;Service&lt;/span&gt; &lt;span class="n"&gt;certsvc&lt;/span&gt;

&lt;span class="c1"&gt;# On a test computer: check a NEW certificate, not an old one&lt;/span&gt;
&lt;span class="n"&gt;certutil&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;verify&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;urlfetch&lt;/span&gt; &lt;span class="n"&gt;test&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="n"&gt;cer&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;In the output, find the &lt;code&gt;Certificate OCSP&lt;/code&gt; section. It should name the responder address and report a verified result.&lt;/p&gt;

&lt;p&gt;The full guide starts on a plain server. It installs the role service and prepares the signing template with the right permissions. It adds the revocation configuration and publishes the address. Then it tests a revoked certificate too, and explains why a certificate you have just revoked can still show as good. It also explains why you should keep your CRLs.&lt;/p&gt;

&lt;p&gt;Read the full guide: &lt;a href="https://learnitlessons.com/articles/ocsp-responder-windows-server/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/ocsp-responder-windows-server/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>pki</category>
      <category>windows</category>
      <category>powershell</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>Ansible cannot find your new role: where roles must live and how to create one</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Tue, 10 Nov 2026 10:10:25 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/ansible-cannot-find-your-new-role-where-roles-must-live-and-how-to-create-one-4klh</link>
      <guid>https://dev.to/vitalii_shumylo/ansible-cannot-find-your-new-role-where-roles-must-live-and-how-to-create-one-4klh</guid>
      <description>&lt;p&gt;You write a role, add it to a playbook and run it. Ansible stops before the first task. It says the role was not found and prints a list of folders. The role is on the disk, so why can Ansible not see it?&lt;/p&gt;

&lt;p&gt;Ansible does not search the whole disk for roles. It checks a short list of places: collections, if you use them, then a &lt;code&gt;roles/&lt;/code&gt; folder next to the playbook, then the folders in &lt;code&gt;roles_path&lt;/code&gt;, and then the folder of the playbook itself. A role that you created in your home folder, or inside another project, is not in any of these places.&lt;/p&gt;

&lt;p&gt;The simple fix is to create the role in the right place from the start. &lt;code&gt;ansible-galaxy&lt;/code&gt; builds the standard folder layout for you, and &lt;code&gt;--init-path&lt;/code&gt; tells it where to put that layout:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;mkdir -p roles
ansible-galaxy role init --init-path roles chrony
ansible-playbook -i inventory site.yml
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Here &lt;code&gt;site.yml&lt;/code&gt; is in the same folder as &lt;code&gt;roles/&lt;/code&gt;, and it lists &lt;code&gt;chrony&lt;/code&gt; under &lt;code&gt;roles:&lt;/code&gt;. If you keep shared roles in another folder, add that folder to &lt;code&gt;roles_path&lt;/code&gt; in &lt;code&gt;ansible.cfg&lt;/code&gt;. Then you do not need to move them.&lt;/p&gt;

&lt;p&gt;The full guide explains what each folder in the layout is for. It builds a small working role that installs chrony, writes its config from a template and restarts the service with a handler. It shows how to change a role setting for one play. It explains why a value in &lt;code&gt;vars/&lt;/code&gt; wins over &lt;code&gt;group_vars&lt;/code&gt;, while a value in &lt;code&gt;defaults/&lt;/code&gt; loses to almost everything. It also compares &lt;code&gt;roles:&lt;/code&gt;, &lt;code&gt;import_role&lt;/code&gt; and &lt;code&gt;include_role&lt;/code&gt;, and shows when each one runs.&lt;/p&gt;

&lt;p&gt;Read the full guide: &lt;a href="https://learnitlessons.com/articles/ansible-roles-explained/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/ansible-roles-explained/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ansible</category>
      <category>devops</category>
      <category>linux</category>
      <category>automation</category>
    </item>
    <item>
      <title>Each user sees a different kiosk desktop: user GPOs follow the account, not the PC</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Fri, 06 Nov 2026 10:04:59 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/each-user-sees-a-different-kiosk-desktop-user-gpos-follow-the-account-not-the-pc-233n</link>
      <guid>https://dev.to/vitalii_shumylo/each-user-sees-a-different-kiosk-desktop-user-gpos-follow-the-account-not-the-pc-233n</guid>
      <description>&lt;p&gt;You build a locked-down desktop for a shared computer, such as a kiosk or a Remote Desktop server. Then people sign in, and each of them sees something different. One person gets a drive mapping from their department, which does not belong on that machine.&lt;/p&gt;

&lt;p&gt;The cause is how Group Policy splits its work. Computer settings come from the OU that holds the computer object. User settings come from the OU that holds the user account. For the user part, the computer that a person signs in to does not matter.&lt;/p&gt;

&lt;p&gt;Loopback processing changes this. You enable it under Computer Configuration in a GPO that applies to the shared computer. After that, the computer's GPOs supply user settings too. There are two modes. Replace ignores the user's usual GPO list and builds the list from the computer's location only. Merge keeps the user's list and processes the computer's list after it, so the computer's GPOs win any conflict. Note that a GPO linked to the domain still applies in replace mode, because the computer is in the domain as well.&lt;/p&gt;

&lt;p&gt;To see which GPOs a user really gets, sign in on the shared computer as that user and run:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight powershell"&gt;&lt;code&gt;&lt;span class="n"&gt;gpupdate&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;/force&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="n"&gt;gpresult&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;/r&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;/scope&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;user&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Some user settings apply only at sign-in, so sign out and in again after the update. Then check the list under "Applied Group Policy Objects". In replace mode, a GPO linked only to the user's OU must not be there.&lt;/p&gt;

&lt;p&gt;The full guide compares the modes on one example with three GPOs. It shows how to create and link the GPO, and why both halves must stay enabled. It covers the permissions the computer account needs, testing with the Group Policy Modeling Wizard, and a common mistake with replace mode.&lt;/p&gt;

&lt;p&gt;Read the full guide: &lt;a href="https://learnitlessons.com/articles/gpo-loopback-processing-explained/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/gpo-loopback-processing-explained/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>activedirectory</category>
      <category>grouppolicy</category>
      <category>windows</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>DHCP failover is up, but only half the clients see your new setting</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Tue, 03 Nov 2026 10:05:39 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/dhcp-failover-is-up-but-only-half-the-clients-see-your-new-setting-df8</link>
      <guid>https://dev.to/vitalii_shumylo/dhcp-failover-is-up-but-only-half-the-clients-see-your-new-setting-df8</guid>
      <description>&lt;p&gt;You have two Windows DHCP servers in a failover relationship. You add a reservation or change a scope option, such as the DNS server, on one of them. After that, some clients get the new value and others keep the old one. It looks random.&lt;/p&gt;

&lt;p&gt;It is not random. In load balance mode, both servers answer clients at the same time. A hash of the client's MAC address decides which server replies, so each server looks after about half of the clients. The two partners keep their lease records in step without your help. Scope settings are a different matter: they stay on the server where you made the change. So only the clients of that server get it.&lt;/p&gt;

&lt;p&gt;The fix is to replicate the relationship after every change. Start it from the server that holds the correct settings, because replication overwrites the configuration on the partner. Then read the value back from the partner:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight powershell"&gt;&lt;code&gt;&lt;span class="n"&gt;Invoke-DhcpServerv4FailoverReplication&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-ComputerName&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;DHCP1&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-Name&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'DHCP1-DHCP2'&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="n"&gt;Get-DhcpServerv4Reservation&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-ComputerName&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;DHCP2&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-ScopeId&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;10.10.10.0&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;For a scope option, use &lt;code&gt;Get-DhcpServerv4OptionValue&lt;/code&gt; with the same &lt;code&gt;-ComputerName&lt;/code&gt; and &lt;code&gt;-ScopeId&lt;/code&gt; values instead.&lt;/p&gt;

&lt;p&gt;The full guide starts earlier. It helps you choose between load balance and hot standby, and it lists the checks to make first, such as the clock difference and TCP port 647. It creates the relationship with one command and confirms that the state is &lt;code&gt;Normal&lt;/code&gt;. Then it stops one server and renews a client to prove that failover really works. It also explains why a hot standby server can stop giving new addresses, and how to let it use the whole pool.&lt;/p&gt;

&lt;p&gt;Read the full guide: &lt;a href="https://learnitlessons.com/articles/dhcp-failover-windows-server/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/dhcp-failover-windows-server/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>windowsserver</category>
      <category>dhcp</category>
      <category>powershell</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>A deleted AD user came back without its groups: the Recycle Bin was off</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Fri, 30 Oct 2026 10:03:52 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/a-deleted-ad-user-came-back-without-its-groups-the-recycle-bin-was-off-21l5</link>
      <guid>https://dev.to/vitalii_shumylo/a-deleted-ad-user-came-back-without-its-groups-the-recycle-bin-was-off-21l5</guid>
      <description>&lt;p&gt;Someone deletes a user account by mistake. You bring it back, but the person still cannot open shared folders or get mail from their groups. The account exists again, yet it is nearly empty.&lt;/p&gt;

&lt;p&gt;This happens when the Active Directory Recycle Bin is not enabled. AD does not erase a deleted object immediately. It keeps a tombstone, but it removes most of the attributes first, and group membership is among them. With the Recycle Bin on, a deleted object keeps every attribute until its lifetime runs out. The feature is off in a new forest. You enable it once for the whole forest, and you cannot disable it later.&lt;/p&gt;

&lt;p&gt;When the feature is on, recovery has two steps. First, look at the deleted object and the OU it came from. Then pass it to &lt;code&gt;Restore-ADObject&lt;/code&gt;:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight powershell"&gt;&lt;code&gt;&lt;span class="n"&gt;Get-ADObject&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-Filter&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'samAccountName -eq "jsmith"'&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-IncludeDeletedObjects&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-Properties&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;LastKnownParent&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="o"&gt;|&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="n"&gt;Format-List&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;Name&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;LastKnownParent&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;Deleted&lt;/span&gt;&lt;span class="w"&gt;

&lt;/span&gt;&lt;span class="n"&gt;Get-ADObject&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-Filter&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'samAccountName -eq "jsmith"'&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-IncludeDeletedObjects&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="o"&gt;|&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="n"&gt;Restore-ADObject&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;code&gt;LastKnownParent&lt;/code&gt; shows the old OU. If that OU was deleted too, the restore fails, because the user has no parent to return to. Restore the OU first, or choose an existing OU with &lt;code&gt;-TargetPath&lt;/code&gt;.&lt;/p&gt;

&lt;p&gt;The full guide shows how to check the feature and enable it, and which rights and forest level you need. It restores a whole OU and its contents in the correct order. It explains what you can still do when the Recycle Bin was off. It also gives one command that protects every OU from accidental deletion.&lt;/p&gt;

&lt;p&gt;Read the full guide: &lt;a href="https://learnitlessons.com/articles/ad-recycle-bin-restore-deleted-objects/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/ad-recycle-bin-restore-deleted-objects/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>activedirectory</category>
      <category>powershell</category>
      <category>windows</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>Promoting a domain controller with PowerShell: the short version</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Thu, 22 Oct 2026 09:51:44 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/promoting-a-domain-controller-with-powershell-the-short-version-an1</link>
      <guid>https://dev.to/vitalii_shumylo/promoting-a-domain-controller-with-powershell-the-short-version-an1</guid>
      <description>&lt;p&gt;Promoting a Windows Server to a domain controller takes two steps, and most failed promotions come from skipping the groundwork before them. The server needs its final name, a static IP address and a DNS setting that points at itself (or at the existing DC) before you start. Renaming a DC afterwards is possible, but it is a fiddly job you can avoid.&lt;/p&gt;

&lt;p&gt;After that, install the role, run the matching &lt;code&gt;Test-&lt;/code&gt; cmdlet, and only then run the real promotion. The test catches the usual problems (DNS, prerequisites, forest level) before anything is changed, so you do not find out halfway through a reboot.&lt;/p&gt;

&lt;p&gt;For a brand-new forest, the first DC looks like this:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight powershell"&gt;&lt;code&gt;&lt;span class="n"&gt;Install-WindowsFeature&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-Name&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;AD-Domain-Services&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-IncludeManagementTools&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="n"&gt;Test-ADDSForestInstallation&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-DomainName&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'corp.example.test'&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-DomainNetbiosName&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'CORP'&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="se"&gt;`
&lt;/span&gt;&lt;span class="w"&gt;    &lt;/span&gt;&lt;span class="nt"&gt;-ForestMode&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'Win2025'&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-DomainMode&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s1"&gt;'Win2025'&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-InstallDns&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;If the test passes, swap &lt;code&gt;Test-ADDSForestInstallation&lt;/code&gt; for &lt;code&gt;Install-ADDSForest&lt;/code&gt;. A second DC uses &lt;code&gt;Install-ADDSDomainController&lt;/code&gt; instead, and the full guide also covers the checks that prove both servers really are working domain controllers: &lt;code&gt;dcdiag&lt;/code&gt;, &lt;code&gt;repadmin /replsummary&lt;/code&gt; and a DNS fix that people often forget on the second DC.&lt;/p&gt;

&lt;p&gt;Read the full guide, with every step and the post-reboot checks: &lt;a href="https://learnitlessons.com/articles/promote-domain-controller-powershell/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/promote-domain-controller-powershell/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;I also keep the lab scripts for this on GitHub: &lt;a href="https://github.com/vitalii-shumylo/it-lab-scripts" rel="noopener noreferrer"&gt;https://github.com/vitalii-shumylo/it-lab-scripts&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>powershell</category>
      <category>windowsserver</category>
      <category>activedirectory</category>
      <category>sysadmin</category>
    </item>
    <item>
      <title>Kerberoasting in plain words: find your exposed service accounts</title>
      <dc:creator>Vitalii Shumylo</dc:creator>
      <pubDate>Thu, 22 Oct 2026 09:51:38 +0000</pubDate>
      <link>https://dev.to/vitalii_shumylo/kerberoasting-in-plain-words-find-your-exposed-service-accounts-5fe9</link>
      <guid>https://dev.to/vitalii_shumylo/kerberoasting-in-plain-words-find-your-exposed-service-accounts-5fe9</guid>
      <description>&lt;p&gt;Kerberoasting is an attack on service accounts, and it works because Kerberos is doing exactly what it was designed to do. Any signed-in domain user can ask a domain controller for a service ticket to any account that has a service principal name (SPN). Part of that ticket is encrypted with a key made from the service account's password, so the attacker takes the ticket away and guesses the password offline. There are no failed logons and no lockouts to notice.&lt;/p&gt;

&lt;p&gt;So the weak point is not Kerberos. It is a human-chosen password on an account with an SPN, made worse when the ticket uses RC4 instead of AES. Computer accounts and group managed service accounts (gMSAs) are safe from this because their passwords are long, random and rotated automatically.&lt;/p&gt;

&lt;p&gt;The first job is to know which ordinary user accounts carry an SPN:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight powershell"&gt;&lt;code&gt;&lt;span class="n"&gt;Get-ADUser&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nt"&gt;-Filter&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="n"&gt;ServicePrincipalName&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="o"&gt;-like&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"*"&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="se"&gt;`&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="nt"&gt;-Properties&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="n"&gt;ServicePrincipalName&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;PasswordLastSet&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;msDS-SupportedEncryptionTypes&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;adminCount&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="o"&gt;|&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="n"&gt;Select-Object&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;Name&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;PasswordLastSet&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;msDS-SupportedEncryptionTypes&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;adminCount&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nx"&gt;ServicePrincipalName&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;An old password, an &lt;code&gt;adminCount&lt;/code&gt; of 1 or no AES setting each push an account to the top of your list. The full guide shows how to move accounts to gMSAs, force AES, and spot an attack by watching event 4769 for RC4 tickets.&lt;/p&gt;

&lt;p&gt;Read the full guide: &lt;a href="https://learnitlessons.com/articles/kerberoasting-explained-prevention/" rel="noopener noreferrer"&gt;https://learnitlessons.com/articles/kerberoasting-explained-prevention/&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;A read-only version of this audit is in my GitHub lab scripts: &lt;a href="https://github.com/vitalii-shumylo/it-lab-scripts" rel="noopener noreferrer"&gt;https://github.com/vitalii-shumylo/it-lab-scripts&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Written with AI assistance and checked by the author.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>security</category>
      <category>activedirectory</category>
      <category>windowsserver</category>
      <category>powershell</category>
    </item>
  </channel>
</rss>
