The Three Tickets That Never Stop
Every Active Directory environment eventually generates the same three categories of pain, over and over: replication silently breaking between domain controllers, Group Policy that Event Viewer swears applied but clearly didn't, and accounts that lock out for reasons nobody can immediately explain. Here's what's actually happening in each, and the specific commands to find and fix the real cause.
Replication Error 8606: Lingering Objects
DC=example,DC=com
Naming Context: CN=Configuration,DC=example,DC=com
Source DC: CN=NTDS Settings,CN=DC02,CN=Servers,...
******* 1 CONSECUTIVE FAILURES since ...
Last error: 8606 (0x21a6):
Insufficient attributes were given to create an object.
What it means: a source DC sends an update for an object that the destination DC already deleted and permanently garbage-collected — because that DC had been offline (or unreachable via replication) for longer than the domain's tombstone lifetime. The destination DC now holds a lingering object: something the rest of the forest agrees no longer exists, but which this one DC still has, and the mismatch blocks replication of that naming context.
Check replication health across all DCs first:
repadmin /showrepl * /csv > replhealth.csv
Open it in Excel/a spreadsheet and filter for non-zero failure counts — this tells you exactly which DC pairs are broken and since when, rather than guessing from one DC's perspective.
The fix is removing the lingering object from the DC that still has it, not fixing "replication" in the abstract:
repadmin /removelingeringobjects <DestinationDC> <SourceDC-GUID> <NamingContext> /advisory_mode
Run with /advisory_mode first — it logs what would be removed (Event ID 1946) without actually deleting anything, so you can verify the list before committing. Microsoft's Lingering Object Liquidator (LoL) automates this across multiple DCs if you have more than a couple affected.
Before chasing lingering objects, rule out error 8524 — a DNS lookup failure that looks similar in symptoms but has nothing to do with tombstones. Destination DCs resolve source DCs by their fully qualified CNAME record; test it directly:
nslookup <source-dc-guid>._msdcs.example.com
If that fails, fix DNS (a missing _msdcs zone, a stale record, or a DC pointing at the wrong DNS server in its own NIC settings) before touching replication metadata at all.
Group Policy That "Applied" But Didn't (Event ID 1058 / 1030)
Event ID: 1058
Source: Group Policy
The processing of Group Policy failed. Windows attempted to read the file
\\example.com\SysVol\example.com\Policies\{31B2F340-016D-11D2-...}\gpt.ini
from a domain controller and was not successful.
Event ID: 1030
Source: Group Policy
Windows cannot query for the list of Group Policy objects. Check the event
log for possible errors. This system will retry group policy processing
at the next processing cycle.
What it means: both point at the client being unable to reach or read the SYSVOL share on a domain controller — usually a DNS resolution problem to the DC, a network path/firewall issue, or (less often) permissions on the SYSVOL folder itself getting corrupted after an FRS-to-DFSR migration that didn't fully complete.
Start with gpresult on the affected machine to see what Group Policy processing actually believes happened:
gpresult /h gpresult.html /f
Then confirm the machine can actually reach and resolve a DC's SYSVOL path directly:
nslookup example.com
Test-Path \\example.com\SysVol\example.com\Policies
If DNS resolves but the path is unreachable, check DFSR replication health for SYSVOL specifically (dfsrdiag replicationstate) — a common cause is DFSR silently stuck in a "journal wrap" or backlog state on one DC, so that DC keeps advertising SYSVOL but serves a stale or inaccessible copy.
If a specific policy setting isn't taking effect even though gpresult shows the GPO as applied, the problem usually isn't connectivity at all — check GPO link order and security filtering: OU-linked GPOs win over domain-linked ones by default, Enforced overrides Block Inheritance, and a GPO can be linked correctly but security-filtered to exclude the exact computer or user you're testing with.
Account Lockouts That Keep Coming Back
What it means: the account is locking out because something is repeatedly authenticating with a stale or wrong password — and it's very rarely the user typing it wrong at their desk. The usual suspects, in order of how often they're the actual cause: a mapped network drive with saved (old) credentials, a scheduled task or Windows service running as that user, and a phone's mail profile still configured with the pre-change password.
Find which DC processed the lockout (in multi-DC environments, this matters — the lockout event lives on the DC that handled the failed authentication attempts, not necessarily the PDC emulator):
Get-ADDomainController -Filter * | ForEach-Object {
Get-WinEvent -ComputerName $_.HostName -FilterHashtable @{
LogName='Security'; Id=4740
} -MaxEvents 5 -ErrorAction SilentlyContinue
}
Event ID 4740 includes a Caller Computer Name field — that's your starting point. Cross-reference it against Event ID 4625 (failed logon) entries on the same DC in the minutes before the lockout to see exactly which service or device is retrying with a bad credential.
# On the DC that logged the 4740, check for preceding 4625 events from the same account
Get-WinEvent -FilterHashtable @{ LogName='Security'; Id=4625 } |
Where-Object { $_.Message -match 'TargetUserName:\s*jdoe' }
Once you've identified the source (a drive mapping, a scheduled task, a service, a phone), update the credential there — clearing the lockout without fixing the source just means it locks out again on the same interval as whatever's retrying.
Quick Reference
| Problem | Real cause | Where to look first |
|---|---|---|
| Replication error 8606 | Lingering object — a DC was offline past the tombstone lifetime | repadmin /showrepl * /csv, then /removelingeringobjects /advisory_mode |
| Replication error 8524 | DNS lookup failure to the source DC's CNAME | nslookup <dc-guid>._msdcs.<domain> |
| Event ID 1058 / 1030 | Client can't reach or read SYSVOL on a DC | gpresult /h, confirm DNS + \\domain\SysVol path is reachable |
| GPO shows "applied" but setting isn't active | Link order or security filtering excludes the target | Check GPO link precedence and security filtering, not just gpresult |
| Repeated account lockouts | A stale credential somewhere is retrying automatically | Event ID 4740's Caller Computer Name, cross-referenced with preceding 4625s |
Sources:
- Troubleshoot replication error 8606 — Microsoft Learn
- Troubleshoot Active Directory replication error 8524 — Microsoft Learn
- Resolution of Active Directory Replication Error 8606 & 1988 — Justin Henson
- Guide to Fixing Common Active Directory Problems — Site24x7
- Active Directory Troubleshooting: Common Errors and Fixes — cagricaliskan.com