AD Replication Error -2146893022: The Target Principal Name Is Incorrect

Fix Active Directory replication error -2146893022 (0x80090322, SEC_E_WRONG_PRINCIPAL): stop the KDC on the destination DC and replicate, check the source DC GUID CNAME and host records, purge tickets, and reset the DC machine password.

Home › Guides › AD replication error -2146893022

Last updated 19 November 2026

Quick answer

Error -2146893022 (0x80090322, SEC_E_WRONG_PRINCIPAL, "The target principal name is incorrect.") means the source DC couldn't decrypt the Kerberos service ticket presented by the destination DC (the one requesting changes). Microsoft's top cause: the KDC that issued the ticket has an old version of the source DC's password. The top fix makes the destination DC get its ticket from another DC:

  1. Disable the KDC service on the destination DC:
    Set-Service -Name KDC -StartupType Disabled
    or sc.exe config KDC start= disabled (sc.exe needs the space after start=).
  2. Restart the destination DC.
  3. Start replication on the destination from the source with AD Sites and Services or repadmin:
    Repadmin /replicate destinationDC sourceDC DN_of_Domain_NC
    Repadmin /replicate ContosoDC2.contoso.com ContosoDC1.contoso.com "DC=contoso,DC=com"
  4. Set the KDC back to automatic and start it:
    Set-Service -Name KDC -StartupType Automatic
    Start-Service -Name KDC
    or sc.exe config KDC start= auto and net start KDC.

If that doesn't fix it, reset the source DC's machine account password with netdom resetpwd (Step 3) and check name resolution (Step 2).

What the errors look like

dcdiag (Replications test):

The replication generated an error (-2146893022):
The target principal name is incorrect.

repadmin /showreps (also /replsum, /showrepl and /syncall):

Last attempt @ <date> <time> failed, result -2146893022 (0x80090322):
The target principal name is incorrect.

repadmin /syncall /Ade:

CALLBACK MESSAGE: Error contacting server CN=NTDS Settings,CN=<server name>,CN=Servers,CN=<site name>,CN=Sites,CN=Configuration,DC=<forest root domain> (network error): -2146893022 (0x80090322):

Replicate now in Active Directory Sites and Services: "The following error occurred during the attempt to contact the domain controller <source DC name>: The target principal name is incorrect". Directory Service events that commonly carry the status:

SourceEventText
NTDS Replication1586The Windows NT 4.0 or earlier replication checkpoint with the PDC emulator master was unsuccessful.
NTDS KCC1925The attempt to establish a replication link for the following writable directory partition failed.
NTDS KCC1308The Knowledge Consistency Checker (KCC) has detected that successive attempts to replicate with the following domain controller has consistently failed.
Microsoft-Windows-ActiveDirectory_DomainService1926The attempt to establish a replication link to a read-only directory partition with the following parameters failed
NTDS Inter-site Messaging1373The Intersite Messaging service could not receive any messages for the following service through the following transport. The query for messages failed.

Why it happens

SEC_E_WRONG_PRINCIPAL isn't an Active Directory error. RPC, Kerberos, SSL, LSA or NTLM can return it. The Kerberos errors Windows maps to it:

Kerberos errorCodeMeaning
KRB_AP_ERR_MODIFIED0x29 / 41Message stream modified (the target can't decrypt the ticket)
KRB_AP_ERR_BADMATCH0x24 / 36Ticket and authenticator don't match
KRB_AP_ERR_NOT_US0x23 / 35The ticket isn't for us

Specific root causes from Microsoft:

Every DC runs the KDC for its domain, and for same-realm requests a destination DC prefers to get tickets from itself. If its own copy of the source DC's password is stale (for example because inbound replication was already failing), its tickets can't be decrypted. That's why stopping the local KDC works: the destination has to get a ticket from another DC. The same KRB_AP_ERR_MODIFIED error on member servers and service accounts usually means a duplicate or misplaced SPN; see KRB_AP_ERR_MODIFIED (duplicate SPN).

Step 1: Check SPNs and the DC account with dcdiag

SPNs can be missing, invalid or duplicated because of replication latency (especially just after promotion) or replication failures. Run this on the console of every source DC that fails outbound replication with this error:

dcdiag /test:CheckSecurityError

To test against a specific partner, the dcdiag reference adds /ReplSource:<SourceDomainController>, which also checks time skew (under 300 seconds) and naming context permissions on that partner.

Step 2: Check that the name resolves to the right DC

The destination DC looks up the objectGUID of the source DC's NTDS Settings object in its own copy of AD, queries DNS for the matching CNAME under _msdcs.<forest root DNS name>, and follows it to the source DC's host A/AAAA record (falling back to the fully qualified name in DNS or a single-label name in WINS). Stale NTDS Settings objects, bad DNS or WINS host records and stale HOSTS entries all send the Kerberos traffic to the wrong target. Microsoft's manual check (no network trace needed):

On the source DC: run IPCONFIG /ALL and note its IP address; run REPADMIN /SHOWREPS and note the DSA Object GUID in the header.

On the destination DC:

  1. IPCONFIG /ALL: note every DNS server it could query.
  2. REPADMIN /SHOWREPS: in Inbound Neighbors, the DSA object GUID listed for the source must match the GUID from the source DC's own header.
  3. IPCONFIG /FLUSHDNS to clear the DNS client cache.
  4. Open %systemroot%\system32\drivers\etc\hosts, remove any entry for the source DC's short or fully qualified name, save, then run Nbtstat -R (upper case R) to refresh the NetBIOS name cache.
  5. Query the GUID CNAME on each DNS server:
    nslookup -type=cname 8a7baee5-cd81-4c8c-9c0f-b10030574016._msdcs.contoso.com 152.45.42.103
    or Resolve-DnsName -Name <objectGUID>._msdcs.<forest root DNS name> -Type CNAME -Server <DNS Server IP>. The address must match the source DC's IP.
  6. Look for duplicate host records for the source DC's FQDN on every DNS server, for example Resolve-DnsName -Name <FQDN of source DC> -Type A_AAAA -Server <DNS Server IP>.
  7. nbtstat -A <IP address returned by nslookup> should return the source DC's name.

If a bad mapping sends the request to a computer that isn't a DC, or to a DC that doesn't have the replication service (UUID E351...) registered with the endpoint mapper, the error becomes 1753: "There are no more endpoints available from the endpoint mapper." Connection failures show up as error 1722.

Step 3: Fix a password mismatch

See which KDC the destination DC is using, close in time to the error, from an elevated prompt:

NLTEST /DSGETDC:<DNS domain of target domain> /kdc

A network trace is the definitive way to see which DC issued the ticket. No Kerberos traffic in a trace can mean the client already has tickets, is getting them from itself, or the trace tool isn't parsing Kerberos. Purge tickets so new ones are requested:

klist purge
klist -li 0x3e7 purge

The first purges the logged-on user's tickets; the second purges the system account's tickets used by AD replication, without a restart. Compare the password attributes of the source DC's computer account (dBCSPwd, UnicodePWD, NtPwdHistory, PwdLastSet, lmPwdHistory) on the source DC and on the KDC the destination uses; obvious version number differences point to the mismatch:

repadmin /showobjmeta <source DC> <DN path of source DC computer account>
repadmin /showobjmeta <KDC selected by destination DC> <DN path of source DC computer account>

To reset a DC's machine account password, run this from an elevated prompt on the console of the DC that needs the reset:

netdom resetpwd /server:<DC to direct password change to> /userd:<user name> /passwordd:<password>

Then start replication again from the destination (repadmin /replicate, as in the quick answer) and confirm the error is gone.

Step 4: Read a network trace or turn up logging

Look up 1925, 1308 and 1926 in the Windows Event ID Lookup, copy the klist, dcdiag and Resolve-DnsName lines from the PowerShell Cheat Sheet, and pull GUIDs and error codes out of repadmin output with the Regex Tester.

Related tools and guides

Look up 1586, 1925, 1308, 1926 and 1373 in the Windows Event ID Lookup, copy klist, dcdiag and Resolve-DnsName lines from the PowerShell Cheat Sheet, extract GUIDs from repadmin output with the Regex Tester, check that the address in a trace belongs to the source DC's subnet with the IPv4 Subnet Calculator or IPv6 Subnet Calculator, and, for a reply from the same subnet, see which adapter vendor answered from its MAC address with the MAC Vendor Lookup. Related guides: KRB_AP_ERR_MODIFIED (duplicate SPN), replication error 1722, replication error 8453 and the trust relationship between this workstation and the primary domain failed.

FAQ

What does AD replication error -2146893022 mean?

It is 0x80090322, SEC_E_WRONG_PRINCIPAL: The target principal name is incorrect. The source DC could not decrypt the Kerberos service ticket that the destination DC presented, so the replication request was rejected.

What is the quickest fix for the target principal name is incorrect?

Microsoft's top resolution: disable the KDC service on the destination DC, restart it, run repadmin /replicate from the source, then set the KDC back to Automatic and start it. If that fails, reset the source DC's machine password with netdom resetpwd.

Why does stopping the KDC on the destination DC help?

A destination DC prefers to get Kerberos tickets from its own KDC. If its copy of the source DC's password is out of date, the source cannot decrypt those tickets. With its KDC stopped, it gets a ticket from another DC that has the current password.

How is -2146893022 related to KRB_AP_ERR_MODIFIED?

Windows maps the Kerberos errors KRB_AP_ERR_MODIFIED (0x29), KRB_AP_ERR_BADMATCH (0x24) and KRB_AP_ERR_NOT_US (0x23) to SEC_E_WRONG_PRINCIPAL. In AD replication the usual causes are a stale DC password on the KDC or a bad name-to-IP mapping that reaches the wrong DC.

How do I check whether DNS points to the wrong domain controller?

On the destination DC, flush the DNS cache, remove HOSTS entries for the source, then resolve the source DC's objectGUID CNAME under _msdcs of the forest root on every configured DNS server. The address must match the source DC's IP, and its DSA Object GUID must match repadmin /showreps on the source.

Want the full kit?

Runbooks, checklists and helpdesk templates: IT Admin Packs on Gumroad.