Skip to content

Est. 2011ยทMicrosoft Partner ID 7033487ยทDelivery under 3 minยทSupport 7 days a week

Your vault is empty.

Free Fix Event ID 7009

Event ID 7009: Exchange setup times out waiting for a service to start

11 min read Updated October 5, 2026 Exchange Server

Fix it now

Event ID 7009 is the Service Control Manager reporting that it stopped waiting for a service to say it had started. During a cumulative update the thing waiting is setup, so setup stops too. Start the named service by hand: the error you get then is the real one, and it is usually a dependency rather than the service itself.

Elevated PowerShell on the Exchange server. Substitute the service named in the 7009 event and your own domain

Get-Service MSExchange* | Where-Object {$_.StartType -eq 'Disabled'}
Start-Service MSExchangeADTopology
sc qc MSExchangeADTopology
nltest /dsgetdc:contoso.com /GC
Test-ServiceHealth
  1. Open Event Viewer, Windows Logs, System, and find the 7009 entry. The service name is the only useful fact in it.
  2. Start that service by hand and read the failure. A dependency that will not start, a missing framework or an unreachable domain controller all surface here in plain words.
  3. Start MSExchangeADTopology first if it is not running. Most other Exchange services wait on it, so a batch of timeouts usually has one cause at the front of the queue.
  4. Put anything found Disabled back with Set-Service <name> -StartupType Automatic, then start it.
  5. Once the services start by hand, resume the update: Setup.exe /IAcceptExchangeServerLicenseTerms_DiagnosticDataON /Mode:Upgrade from an elevated Command Prompt.

Exchange setup is resumable. Running the same package again in upgrade mode continues from where it stopped, and is the supported way to recover a run that failed part way.

If the services start and setup completes, stop here. If they will not start, or start too slowly, the next section takes the two apart.

Why it happens

The Service Control Manager gives every service a fixed window to report that it has reached the running state. That window is machine-wide and identical for a Windows service, a backup agent and an Exchange service. When it closes with no answer, the SCM writes the event and reports failure to whoever asked – which during a cumulative update is setup, so the update stops with it.

Microsoft’s own article on this timer covers events 7000 and 7011 and documents the value that governs it: ServicesPipeTimeout, a DWORD under HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control, which the article raises to 60000 milliseconds. It does not publish a default, so do not repeat one. The number that matters is in the event text itself, which records the interval that actually elapsed.

Exchange makes the timer visible because its services form a chain. MSExchangeADTopology runs first: it discovers the domain controllers and global catalogs in the server’s Active Directory site and hands that topology to everything else. Most other Exchange services depend on it. If the directory is slow, unreachable or has no global catalog serving the site, the topology service waits, everything behind it queues, and the timeouts arrive together.

A cumulative update raises the odds because a CU is a full reinstall. Setup stops the services, replaces the binaries and starts everything again on a server whose disk has just been hammered. Slow storage, an antivirus product without the Exchange exclusions, and a reboot still pending from the last patch round all stretch the start past the window. 0x800F0922 belongs to a different layer entirely: it is CBS_E_INSTALLERS_FAILED, “Processing advanced installers and generic commands failed”, which is Windows component servicing reporting its own failure rather than a service timing out.

An Exchange service has been set to Disabled

You have this one if Get-Service reports the named service with a StartType of Disabled rather than Automatic.

  1. Put it back and start it: Set-Service <name> -StartupType Automatic then Start-Service <name>.
  2. Find out what did it. Run gpresult /h C:\gp.html and look under Computer Configuration, Windows Settings, Security Settings, System Services.
  3. If a policy is the source, exclude the Exchange servers from it before you rerun setup, or the next CU fails identically.
  4. Rerun setup in upgrade mode from an elevated prompt.

Backup and antivirus agents sometimes stop Exchange services for the duration of a job and do not put them back when the job aborts. Check the agent’s log for the same timestamp.

The topology service cannot find a directory server

You have this one if The 7009 names MSExchangeADTopology, or names a group of services that all depend on it, with MSExchange ADAccess entries in the Application log at the same moment.

  1. Ask Windows to find a global catalog from that server: nltest /dsgetdc:contoso.com /GC, and add /FORCE to make it query DNS rather than the cache.
  2. Confirm the server’s DNS settings list only domain controllers hosting the Active Directory zone. A public resolver anywhere in the list breaks locator queries intermittently.
  3. Confirm the site Windows thinks the server is in: nltest /dsgetsite.
  4. Check the paths are open to the controllers: LDAP on 389, global catalog on 3268, Kerberos on 88 and the RPC endpoint mapper on 135 with its dynamic range.
  5. Start MSExchangeADTopology by hand, let the dependent services follow, then resume setup.

The Windows servicing part of the update failed

You have this one if 0x800F0922 in the setup log or Windows Update history, and the update stops at or just before the service-start step.

  1. Read C:\ExchangeSetupLogs\ExchangeSetup.log from the bottom and note the last step attempted.
  2. Look underneath it in C:\Windows\Logs\CBS\CBS.log and C:\Windows\Logs\DISM\dism.log, which is where the servicing failure is actually recorded.
  3. Repair the component store: sfc /scannow, then DISM /Online /Cleanup-Image /RestoreHealth.
  4. Confirm the installed .NET Framework matches what your Exchange build requires. Exchange Server SE requires .NET Framework 4.8 or 4.8.1.
  5. Reboot and rerun setup in upgrade mode.

A .NET Framework version newer than the build expects fails as reliably as one that is older. Check the requirement rather than assuming newer is safer.

The service does start, but not inside the window

You have this one if By the time you log on to investigate, the service is running normally and nothing looks wrong.

  1. Confirm it is slowness rather than a dependency failure: start the service by hand and time it.
  2. Apply Microsoft’s published antivirus exclusions for Exchange servers before the next attempt. This is the single change that fixes most of these.
  3. Check storage latency on the database and log volumes during the start, not afterwards.
  4. If you need the update finished tonight, raise the machine-wide timeout: a DWORD named ServicesPipeTimeout under HKLM\SYSTEM\CurrentControlSet\Control, set to 60000, then restart the server.

Raising the timeout is a workaround, not a fix. It is machine-wide, it affects every service on the box, and it hides slowness rather than removing it. Take the registry value out again once the underlying cause is dealt with.

Full reference

Reading the pattern before touching anything

What the logs show Where the fault actually is
7009 naming MSExchangeADTopology No usable domain controller or global catalog for this server’s Active Directory site
7009 naming a transport or mailbox service while ADTopology runs That service failed to initialise. Read its own Application log entries
7022 rather than 7009 The service started and then stopped responding. Look at storage latency and scanners
0x800F0922 in the setup log Windows component servicing failed. The detail is in CBS.log, not in the Exchange log
The service shows as Disabled Policy, a backup agent or an aborted setup turned it off
Several services time out together One dependency at the front of the chain. Fix that, not the six behind it

Commands worth having open

Command What it answers
sc qc <service> What the service depends on and how it is configured to start
sc query <service> Its current state
Get-Service MSExchange* | Where-Object {$_.StartType -eq 'Disabled'} Which Exchange services have been switched off
Set-Service <name> -StartupType Automatic Puts one back
nltest /dsgetdc:<domain> /GC Whether Windows can locate a global catalog from this server
nltest /dsgetsite Which Active Directory site Windows thinks the server is in
Test-ServiceHealth Which required Exchange services are not running for the installed roles
Get-ExchangeServer | Format-List Name,Edition,AdminDisplayVersion The build, once the update finally lands

The registry value, and its price

Item Value
Key HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control
Value name ServicesPipeTimeout
Type DWORD
Data Microsoft uses 60000 (milliseconds)
Scope Every service on the machine, not just Exchange

Export the Control key before you edit it and restart the server afterwards. This is a production mail server and the key is not a place to improvise.

Resuming a cumulative update that stopped

Exchange setup is resumable and rerunning it is the supported recovery. Put database availability group members back into maintenance mode, open an elevated Command Prompt, and run Setup.exe /IAcceptExchangeServerLicenseTerms_DiagnosticDataON /Mode:Upgrade again. Setup continues from the step that failed. What is not supported is uninstalling Exchange to tidy up, or installing an older cumulative update over a newer one.

If setup’s readiness checks fail, they can be corrected and retried without leaving setup – but where the correction needs a restart, Microsoft’s instruction is to start the whole run again rather than resuming after the reboot.

Two events with no published description

Neither Event ID 7009 nor Event ID 7022 has a Microsoft page describing it, so treat the numbers as labels rather than as diagnoses. What you can rely on is the event’s own text: 7009 names the service and the interval waited, and 7022 names a service that reached a state and then stopped answering. Everything useful comes from starting the service by hand and reading what it says, not from decoding the event ID.

Every code this article covers

Code What it points at Source
Event ID 7009 The Service Control Manager stopped waiting for the named service to report that it had started. Microsoft publishes no description page; the event text carries the service name and the interval not published by the vendor
Event ID 7011 A service did not respond inside the Service Control Manager timeout. Microsoft’s article on this timeout documents ServicesPipeTimeout as the setting that governs it Microsoft Learn
0x800F0922 CBS_E_INSTALLERS_FAILED: processing advanced installers and generic commands failed. A Windows component servicing failure, not a service timeout Microsoft Learn
Event ID 7022 The Service Control Manager recording that a service started and then stopped responding. No published description; read the event text not published by the vendor

Confirm the fix worked

  1. Test-ServiceHealth reports no required services missing for the roles installed.
  2. Get-ExchangeServer | Format-List Name,Edition,AdminDisplayVersion shows the build you were installing.
  3. C:\ExchangeSetupLogs\ExchangeSetup.log ends in success rather than a rollback.
  4. Reboot once and confirm every Exchange service reaches Running with no new 7009 or 7022 entries in the System log.
  5. If you added ServicesPipeTimeout, confirm the services still start after you remove it again.

Questions people ask about this

Does this cost anything to fix?

No. It is a configuration and timing problem on a server you already license. Money only enters the conversation if you discover the installed version is out of support and no forward cumulative update exists.

Can I just raise the timeout and move on?

You can get the update finished that way, and sometimes that is the right call at 2am. But the setting is machine-wide and it masks the cause. Come back to the storage latency, the missing antivirus exclusions or the unreachable domain controller once the window closes, and take the value out.

Is it safe to rerun a cumulative update that failed halfway?

Yes, and it is the supported approach. Run the same package again in upgrade mode from an elevated prompt and setup continues from the failed step. What is not safe is uninstalling Exchange to clean up, or going backwards to an older CU.

Should the server be in maintenance mode before a CU?

On a database availability group member, yes: Microsoft’s cumulative update guidance points at the DAG maintenance procedure explicitly. Draining databases and taking the server out of the cluster’s way removes a whole class of timing failures, this one included.

Why do six services time out at once?

Because most Exchange services depend on the Active Directory Topology service. When it cannot get a topology, everything behind it waits and then times out together. Fix the one at the front of the chain and the other five never appear.

Related error codes

Was this article helpful?

Your feedback helps us improve our documentation.

Related articles

Free Fix 451 4.4.0 DNS query failed: Exchange cannot resolve the next hop License Error HCW8078: the hybrid migration endpoint could not be created by the wizard License Error StorageTransientException: the mailbox is busy or over quota during a move Free Fix Error 1618 and a half finished Exchange CU: setup will not start again
โ† Back to Knowledge Base