Troubleshooting Amazon SMS - Amazon Server Migration Service
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China.

Product update

We recommend Amazon Application Migration Service (Amazon MGN) as the primary migration service for lift-and-shift migrations. If Amazon MGN is unavailable in a specific Amazon Region, you can use the Amazon SMS APIs through March 2023.

Troubleshooting Amazon SMS

The following information can help you troubleshoot issues with errors that you might encounter when using Amazon SMS. Before using these procedures, confirm that your SMS setup and the server you are trying to migrate meet the requirements in Requirements for Amazon Server Migration Service.

Log files for the connector

The Server Migration Connector provides log files that you can use to troubleshoot replication jobs that fail prior to completing the upload to Amazon S3. Use the following procedure to download the connector log files.

To download the connector log files

  1. In a web browser, enter the IP address of the connector VM.

  2. Log in to the connector.

  3. Verify that the connector passes all checks.

  4. Under Support Links, choose Download Log Bundle.

  5. Extract the files in the log bundle.

The following connector log files are included in the log bundle:

  • connector.log – Check for connector configuration issues.

  • connectorsetup.log – Check for detailed information about the initial configuration.

  • frontend.log – Check for issues with connectivity to Amazon endpoints.

  • metrics.log – Check the throughput statistics and upload speeds (see UploadStats).

  • netstat.log – Check for network packet errors.

  • poller.log – Confirm database polling activity.

  • sms-replication-poller-log – Review activity from validation of the replication job through the disk is uploaded to Amazon S3. For example, you can verify upload progress as a percentage and review the start and end of each phase of the replication job.

Failure when registering the connector

If you encounter an issue registering the connector, contact

Certificate error when uploading a VM to Amazon S3

The connector may fail to replicate your VM because the VM is on an ESXi host with an SSL certificate problem. If this occurs, you see the following error message displayed in the Latest run's status message section: "ServerError: Failed to upload base disk(s) to S3. Please try again. If this problem persists, please contact Amazon Support: vSphere certificate hostname mismatch: Certificate for <> doesn't match any of the subject alternative names: [localhost.localdomain]."

You can override this ESXi host certificate problem by completing the following tasks:

Upgrade your connector

This section is for customers who are manually upgrading the connector. If you have previously configured automatic upgrades, skip these steps and continue to Re-register your connector.

To upgrade your connector

  1. Open the connector console.

  2. Log in to the connector.

  3. Choose Upgrade.

  4. Wait for the connector to finish upgrading to version or later.

Re-register your connector

This section applies to all customers encountering the certificate mismatch problem.

To re-register your connector

  1. Open the connector console.

  2. Log in to the connector.

  3. In the General Health section, check that the connector version is or later.

  4. Choose Edit Amazon Server Migration Service Settings.

  5. On the Setup page, for Amazon Region, select the desired region from the list. For Amazon Credentials, enter the IAM access key and secret key that you created in Step 2 of the setup guide. Choose Next.

  6. On the vCenter Service Account page, enter the vCenter hostname, user name, and password that you created in Step 3 of the setup guide.

  7. Select the Ignore hostname mismatch and expiration errors for vCenter and ESXi certificates check box. Choose Next.

  8. Complete registration and view the connector configuration dashboard.

  9. Use the Amazon SMS CLI or API to delete and restart your stuck replication jobs.

Server Migration Connector fails to connect to Amazon with error "PKIX path building failed"

In some customer environments, secure network traffic is proxied through a certificate re-signing mechanism for auditing and management purposes. This can cause your Amazon credentials to fail when the connector attempts to contact Amazon SMS. The error message contains "PKIX path building failed," indicating that an invalid certificate was presented.

For the connector to work in such an environment, the re-signing certificate (a user certificate that your organization trusts and uses to sign outbound packets) must be added to the connector's trust store, as described in the following steps.

To add the re-signing certificate to the connector trust store

  1. On your connector system, disable the FreeBSD packet filter and enable SSH with the following commands:

    sudo service pf stop sudo service sshd onestart
  2. Copy your user certificate to the connector by a method such as the following:

    scp userCertFile ec2-user@
  3. Add the user certificate to the trust store:

    keytool -importcert -keystore /usr/local/amazon/connector/config/jetty/trustStore -storepass AwScOnNeCtOr -file /tmp/userCertFileName -alias userCertName
  4. Restart services using the following command (part of Amazon Management Portal for vCenter):

    sudo setup.rb

    Select option 3 and type "yes".

  5. Re-enable the packet filter:

    sudo service pf start

This CA Root certificate is not trusted

When you access the IP address of a virtual machine that you installed on-premises, you may receive the following message:

This CA Root certificate is not trusted. To enable trust,
install this certificate in the Trusted Root Certifications
Authorities store.

You can safely ignore this message.

Replication run fails during the preparing stage

In some cases, Amazon SMS allows a replication job to continue scheduling incremental replication runs even when the latest replication run has failed. When the maximum allowed number of consecutive failures is reached, the default behavior for a replication job is to be paused. The job can be resumed within four days, after which it is deleted. In such cases, the Amazon EBS snapshots from the latest replication run are shared with the customer account, and a status message for the failed replication run is sent. The message contains the snapshot IDs and states the reason for the failure. A typical status message resembles the following:

EBS snapshot(s) created with snapshot ID(s): snap-12345678abcdefgh. Another run  
has been scheduled after the last run failed due to an import failure. 2 re-try run(s) 
remaining before the job will be failed.

The reason for replication-run failures (including first-boot failures) often correlates closely with failures observed when Amazon EC2 VM Import/Export is used for VM migrations. For more information, see Troubleshooting VM Import/Export.

If you need further help with resolving a problem, contact Amazon Web Services Support. EBS snapshots generated during a failed migration are shared with your account, and the snapshot IDs are included in the status message for the replication job. Be sure to have these details available when you contact Amazon Support.

Replicated AMI doesn't support some instance types for launch

Some instances require ENA support. If the migration does not enable ENA support, then the replicated AMI does not allow you to launch instances that require ENA support.

Verify that ENA is enabled. For more information, see Enabling Enhanced Networking on Windows or Enabling Enhanced Networking on Linux in the Amazon EC2 documentation.

ServerError: Failure to upload base disk(s) to Amazon S3

Possible causes

  • The VMDK is not snapshottable or the VM has mounted ISOs.

  • The connection to the hypervisor (Hyper-V or ESXi host) timed out while the connector uploads buffered data to Amazon S3.

  • Maintenance is being performed while the replication job uploads disks to Amazon S3.

  • There is a compression issue with the virtual disk.

  • There is a validation error with the hypervisor certificate.

  • The status of the connector is Unhealthy.

  • The connector cannot reach Amazon endpoints.

ServerError: Failed to validate replication job

Possible causes

  • There is a change in the virtual machine path.

  • There is a change in IAM permissions.

  • There is a change in user or account permissions for the virtual environment.

  • There is a configuration issue with WinRM (Hyper-V).

  • There is a DNS resolution failure.

  • There is an NTP configuration error on the connector VM.

An internal error occurred. Confirm that your Amazon credentials and VM Manager credentials are correct.

Possible causes

  • The IAM permissions are not sufficient to complete the connector setup.

  • The user or account permissions for the virtual environment are not sufficient.

  • There are issues with the IAM roles for Amazon SMS.

  • There are missing prerequisites.

  • The VM environment is not prepared.

  • Special characters were used when setting up the connector (Hyper-V).

Snapshot-related errors (VMware)

Possible causes

  • The VMDK is configured as an independent disk.

  • The ESXi host cannot take a snapshot.

  • The VMDK is locked.

  • The snapshot chain is broken. Ensure that no snapshots are taken between replication runs, either manually or by third-party software.

  • A previous replication run did not consolidate snapshots.

Checkpoint errors (Hyper-V)

Possible causes

  • The VM has existing checkpoints.

  • There are checkpoints created manually or by third-party software.

  • The VHD or VHDX is locked.

  • The Hyper-V host is unable to create a checkpoint.

Incremental replication delta exceeds 1 TB

The connector is designed to handle frequent replication with small deltas. The connector does not support deltas larger than 1 TB. If you do not replicate on a regular basis, the delta can exceed this limit and the replication run fails.

To prevent this issue, set up frequent incremental replication runs. If you cannot replicate frequently, you can increase the delta upload limit. For example, run the following commands on the connector to increase the part size of S3 uploads from 25 MB to 100 MB. When prompted, select option 3.

sudo sms-connector-config -set slotSizeMB 100 sudo setup.rb

Increasing the upload limit impacts the performance and memory usages of the connector. Do not increase the upload limit while the connector is uploading multiple deltas.