Hi Pisek Siripoka,
Thank you for the detailed follow-up and for holding off on changes. Your checks are useful, but error 8240 alone does not confirm that the container is missing. The command may also be using an incorrect or unavailable search base. Microsoft’s documented search uses ncobj:domain: so you do not need to enter the domain naming context manually.
Answers to your questions
- Read-only search: Run this command against each writable domain controller:
repadmin /showattr <DC-FQDN> ncobj:domain: /filter:"(objectClass=msDS-PasswordSettingsContainer)" /subtree
It searches the domain naming context, including for a container that may have been moved. If you use an explicit naming context instead, first verify it against the DC’s defaultNamingContext. An 8240 result from the earlier command is not, by itself, proof that the container is absent throughout the domain.
- If the container is confirmed absent and cannot be restored: Microsoft documents resetting the applicable DomainUpdates state and then running adprep /domainprep. This is a specific recovery procedure; do not run /domainprep alone as a substitute. The procedure is performed on the Infrastructure Master by a Domain Admin. See Microsoft’s KB guidance for a moved or missing Password Settings Container and the Adprep documentation.
- Production impact: The Microsoft article documents the recovery steps, but it does not provide a production success rate or guarantee uninterrupted authentication. It does not instruct you to restart or demote the Infrastructure Master as part of those steps, but that does not guarantee there will be no service impact.
- Identifying the operation marker: Revision 16 is consistent with Windows Server 2016 domain preparation, but it does not identify which operation marker to remove. The KB gives 71482d49-8870-4cb3-a438-b6fc9ec35d70 as an example and says to remove the marker for the operation that needs to be rerun. Do not delete all markers or select one based only on revision 16; confirm the exact object and its full distinguished name against the KB procedure before making a change. See Microsoft’s domain-wide updates reference.
Recommended next steps
First, run the read-only search against every writable DC. If it returns the container, record its distinguished name and follow the KB guidance to move it to the expected location under CN=System.
If no result appears on any DC, check replication health before considering a repair:
repadmin /replsummary
dcdiag /e /c /v
Before making directory changes, confirm you have recent, valid system-state backups and a recovery plan. Microsoft’s AD forest recovery guidance covers system-state backups. Also record the current revision value and the objects under CN=Operations,CN=DomainUpdates,CN=System on the Infrastructure Master.
If the applicable marker is not clear from the KB and those records, pause and have an AD DS specialist review it before deleting anything or resetting the revision. If the container is confirmed absent, replication is healthy, and the correct marker is verified, schedule the KB procedure as a controlled change on the Infrastructure Master. Afterward, review ADPrep.log and confirm the container exists under CN=System and has replicated to all writable DCs.
Thank you again for gathering these results carefully. If you share the search results and replication summary—with domain and server names redacted—we can review them before you decide whether to proceed.