About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
A recovery plan test in VMware Site Recovery Manager (SRM) or VMware Live Site Recovery stops at the power-on step for one VM:
Timed out waiting for VMware tools after [300 or 900] seconds.
KB 430479 describes it: the VMs "successfully execute the "Power On" step but fail shortly after. The recovery plan terminates with a timeout error, preventing the VM from running at the recovery site." So the VM powered on. What SRM did not get was a signal from inside the guest. The fix depends on why that signal never came. The guest may have hung during boot, or been slowed by load, or been fine but simply needed more time.
What the error means
After power-on, SRM waits for VMware Tools in the guest to report. KB 430479: "Because the OS fails to initialize its kernel or services, the VMware Tools daemon (vmtoolsd) never starts. Without vmtoolsd, the GuestRPC channel used to communicate the "OS Family" and "Heartbeat" to SRM is never established." SRM needs this before it can run IP customization or later steps.
SRM's own log, vmware-dr.log, shows the wait and the failure. KB 430479:
Waiting for OS family to become known from VM [vim.VirtualMachine:##:vm-######]; timeout: 900 seconds
Wait for VM Tools in VM VM_NAME[vm-######] failed: (dr.recovery.fault.ToolsTimedout)
The VM's own vmware.log shows the guest side. A guest that never loads Tools leaves "Tools: Delaying state change request to state 3." and then, after ten minutes, "Tools: No activity for 10 minutes, resetting Tools version." KB 430479: "If the OS hangs during boot, we will see this "Delaying" message, and the state change will never actually finish."
On the 300-second form, KB 430479 adds: "If the error occurs at 300 seconds, it typically indicates a delay in VMware Tools initialization during isolated network tests, often before the longer 900-second watchdog timer is reached."
Likely causes
| Cause | Sign | Same error text? | Source |
|---|---|---|---|
| Guest hung during boot | Console stuck; no Tools line in vmware.log after the Delaying line | Yes | KB 430479 |
| Load from many VMs starting together | Database and app servers in one plan; slow boots | Yes | KB 430479 |
| Healthy guest that needs longer than the timeout | Guest boots, but after the wait expires | Yes | KB 430479 |
| Isolated test network constraints | 300-second failures in bubble tests | Yes | KB 430479 |
| VGAuth service not running (IP customization) | VM powers off; "vix error codes = (1, 0)" | No, a related fault | KB 447641 |
| IP customization script cannot reach the host | Failover plan stuck at 96%; ports 902 and 443 are needed for the script upload | No, a related fault | KB 443108 |
KB 430479 lists the scenarios: "High-I/O Contention: Heavy consolidated loads, particularly with high-I/O virtual machines (e.g., database servers), can cause storage and synchronization contention that delays VMware Tools initialization." and "OS Boot Hangs: The Guest OS is stuck (e.g., Windows "Getting devices ready" or Linux fsck prompt)."
Prove the cause before changing anything
These steps only read, apart from running a test recovery, which uses the recovery site's resources.
1. Note the timeout and the step. Is it 300 or 900 seconds? Which VMs? Do they share a priority group with heavy database servers?
2. Watch the console during a test. KB 430479: "Perform a Test Recovery and immediately open the VM Console." It lists what to look for. Windows: "Stuck on update screens or "Getting devices ready."" Linux: "Stuck at a fsck prompt or emergency mode."
3. Read vmware-dr.log from the SRM server. KB 312791 gives the way to get the server logs: in the Site Recovery home tab, select the site pair, View Details, then Actions in the Site Recovery box, select a server and click Export Logs, then Download. On the appliance, KB 312791 also describes a log bundle script, dr-support-linux.sh in /opt/vmware/dr/bin (/opt/vmware/vmware-dr/bin on Live Recovery combined appliance 9.0.3 and above), with output in /var/log/vmware/Support. Find the "Power on succeeded" line and the "Waiting for OS family" line for the VM. The time between them and the failure line tells you whether SRM waited the full period.
4. Read the VM's vmware.log. It sits in the VM's folder on the recovery datastore. KB 323131 shows the path form /vmfs/volumes/VM_Directory/vmware.log. KB 430479 logs "Tools: Delaying state change request to state 3." in the same second as power-on, so that line alone proves nothing: it appears for slow guests too. What separates the cases is what follows. A guest that never loads Tools adds no later Tools line, and after 600 seconds the host logs "Tools: No activity for 10 minutes, resetting Tools version." That line comes ten minutes after power-on, after a 300-second SRM timeout has already fired, so for 300-second failures read vmware.log again about ten minutes later. KB 430479 does not show the line a guest writes when Tools does come up, so use the console as the second check: a guest that reaches its login screen after the SRM timeout was slow, not stuck.
5. If the VM powers off during customization, read the guest log. KB 447641 gives the path C:\Windows\Temp\VMware-imc\srmDeployGuestAlias.log, where VGAuthCLI.exe calls end with "Returned: -1". Its cause: "The root cause of this failure is that the "VMware Alias Manager and Ticket Service" (VGAuthService) is NOT running inside the target VM at the moment of power-on during the SRM/VLSR test." To check, power on the failed test VM, run services.msc and find that service. KB 447641: "The service will typically be found stopped, with the Startup Type configured as Manual or Disabled."
6. If a failover plan hangs at a percentage, test the ports. This is a related fault, not the Tools timeout error. KB 443108 describes a Failover Recovery Plan "stuck at percentage (e.g., 96%)" in Recent Tasks, and for 96% says it "almost always indicates that the VMs have successfully powered on ... but VLSR is waiting for a VMware Tools heartbeat or an IP customization script to complete." The ports matter only when IP customization is part of the wait. KB 443108: "IP customization requires uploading a script from VLSR to the local ESX host and further into the guest OS. Ports 902 & 443 are required to be open for this." It prints:
nc -z {ESX} 902
nc -z {ESX} 443
Example (illustrative): a plan has a database server and six application servers in one priority group. In the test, two application servers fail after 300 seconds. On the console both are still booting, slowly but cleanly. Both reach the login screen about a minute after SRM gave up, and vmware.log has no "No activity for 10 minutes" line ten minutes after power-on. That is load, not a broken guest. Splitting the database server into its own plan, or raising the wait for those VMs, fits KB 430479's guidance.
Fix
Guest hung during boot. KB 430479: "Address the specific OS-level failure (e.g., repairing the boot loader, fixing driver conflicts, or clearing inconsistent filesystems)." Fix the protected (source) VM so the replica inherits the fix.
Load from simultaneous power-ons. KB 430479: "Segregate high-I/O database systems and standard application servers into discrete Protection Groups and Recovery Plans to reduce simultaneous resource contention during recovery." Risk: changing protection groups changes what a real failover recovers and in what order. Agree the plan with the application owners.
Healthy but slow guest. KB 430479: "In the SRM Interface, navigate to Recovery Plan > Recovery Steps. Right-click the affected VM and select Configure Recovery. Navigate to Startup Action > Power On > Wait for VMware tools. Increase the value from the default (300 seconds) to 900 seconds (or 1800 seconds for extreme cases)." Risk: a longer wait makes a real failover take longer when a guest truly is stuck. Way back: set the value back to 300 seconds in the same Configure Recovery dialog.
Unchecking the wait. KB 430479: "if IP customization is not required for a specific VM, you may uncheck "Wait for VMware tools" (this will prevent IP customization)." KB 394491 shows the other side: "The IP customization steps are skipped because the “Wait for VMware Tools” option is disabled". Risk: a VM that needs a new IP address at the recovery site will come up with its old one. Way back: tick "Wait for VMware tools" again.
VGAuth service stopped. Fix it on the source VM, not the test copy. KB 447641: "Manually starting the service or changing settings on this target VM will only fix the current temporary instance." On the source VM: "Right-click the service, select Properties, and change the Startup type to Automatic." Then "Allow the replication mechanism ... to complete at least one successful delta sync cycle." Run the Clean up task and test again.
Failover plan stuck at 96%. KB 443108: "Ensure VMware Tools is running on failed over guest. Manually start tools guest on if possible." If the nc -z test above fails for port 902 or 443, those ports must be open between VLSR and the ESXi host, because IP customization needs them. Risk: opening them is a firewall rule change. Open only those two ports for that path, and way back is to remove the rule when it is no longer needed.
Failover plan stuck at 84%. KB 443108: in the Site Recovery UI, select Recovery Plans, find the stuck plan and cancel the "In Progress" or "Incomplete" Recovery Plan. Then check that the status changes to "Recovery Interrupted" or "Completed." Risk: cancelling a plan interrupts it. Re-run it afterwards.
Other power-on failures. If the error is "No host is compatible with the virtual machine", KB 396530 points to an EVC mode the recovery hosts do not have; its fix needs the VM powered off. Disruptive step: KB 396530: "The EVC configuration will need to be disabled for the affected VM. The VM will need to be powered off", so the protected VM has downtime. Way back: re-apply the VM's EVC setting after the test. Do not power on placeholder VMs by hand. KB 425594: "An SRM Placeholder VM is not a functional virtual machine; it is a "shell" containing only metadata files (.vmx, .vmsd)."
Confirm the fix
- A test recovery completes the power-on step for every VM.
- vmware-dr.log shows no "ToolsTimedout" fault.
- Guests boot to the login screen on the console.
- IP customization ran where it is configured.
- The Clean up task completes after the test.
What to watch afterwards
- Guest patching on protected VMs. Update screens at boot cause hangs (KB 430479).
- New heavy VMs added to existing plans.
- VMware Tools on protected VMs. KB 443108: "Whenever possible, always keep VMware Tools uptodate on protected VM and synced."
- Service changes on source VMs that turn VGAuthService to Manual.
Read next
Related articles you may want next.
- Backup & RecoverySRM Test Would Use the Recovery Port Group, Not the Isolated NetworkAn SRM recovery plan test would attach VMs to a DR port group, not the isolated test network. Trace which setting wins: the plan's or the site mapping.Read article
- VMwareVMware Tools Not Running or Out of Date: What It BreaksRead VMware Tools running and version status correctly, see which guest operations depend on it, and pick the maintenance route for Tools or open-vm-tools.Read article
- Backup & RecoveryHow to implement and test VMware-to-Azure disaster recoveryBuild a modernised Azure Site Recovery workflow with isolated recovery drills and a reviewed return path.Read article
Risks and limits
- KB 443108 is written for VMware Live Recovery 9.0 to 9.0.5.
- KB 430479 times are from one example log; your timestamps will differ.
- A test recovery uses the recovery site's resources. Run it in a window agreed with whoever runs that site.
NetLeverage runs its own hosting on VMware vSphere, vCenter and ESXi. Talk to an engineer.
Sources
- Error: "Timed out waiting for VMware tools after [300/900] seconds" during SRM Test Recovery (KB 430479) — Broadcom, Updated On 10-30-2026, accessed 17 November 2026
- SRM/VLSR IP customization fails with exit code -1 due to inactive VGAuth service on target VM (KB 447641) — Broadcom, Updated On 07-10-2026, accessed 17 November 2026
- IP Customization tasks not initiated for specific virtual machines in a recovery plan (KB 394491) — Broadcom, Updated On 04-13-2026, accessed 17 November 2026
- Failover Recovery Plan stuck or incomplete in VMware Live Recovery (KB 443108) — Broadcom, Updated On 09-14-2026, accessed 17 November 2026
- SRM Recovery Plan Test fails with error: No host is compatible with the virtual machine (KB 396530) — Broadcom, Updated On 05-06-2025, accessed 17 November 2026
- Manual Power-on of SRM Placeholder VM fails with "Incompatible version -1 (expect 2). Module NamespaceMgr power on failed" (KB 425594) — Broadcom, Updated On 09-09-2026, accessed 17 November 2026
- Collecting diagnostic information for VMware Live Site Recovery or VMware Site Recovery Manager (KB 312791) — Broadcom, Updated On 10-03-2026, accessed 17 November 2026
- Extending a virtual disk fails with the error:msg.disklib.UNSUPPORTEDFEATURE (KB 323131) — Broadcom, Updated On 02-24-2026, accessed 17 November 2026
This article was originally published on .