Cluster Image update via PowerCLI cmdlet Import-LcmClusterDesiredState fails with "Record already exists"
Article ID: 425868
Updated On:
Products
Issue/Introduction
When attempting to update the current vSphere Lifecycle Manager (VLCM) image for a cluster using PowerCLI, it fails with a error similar to the one below:
PS C:\GSlabs\Scripts> Get-Cluster -Name ### | Import-LcmClusterDesiredState -LocalSpecLocation '<path\to\json-file>' -Confirm:$false
Import-LcmClusterDesiredState : <timestamp> Import-LcmClusterDesiredState 500: Internal server error calling EsxSettingsClustersSoftwareDraftsImportSoftwareSpec: {"messages":[{"args":["import_software_spec","com.vmware.vapi.std.errors.already_exists"],"default_message":"Unexpected error for method: import_software_spec -> com.vmware.vapi.std.errors.already_exists","id":"vapi.bindings.method.impl.unexpected"},{"args":[],"default_message":"Record already exists.","id":"com.vmware.vcIntegrity.lifecycle.drafts.AlreadyExists"}]}
At line:2 char:30
+ ... <clustername> | Import-LcmClusterDesiredState -LocalSpecLocation '### ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : NotSpecified: (:) [Import-LcmClusterDesiredState], CisException
+ FullyQualifiedErrorId : ViCore.Impl.V1.Service.LcmService.ImportDesireState.Error,VMware.VimAutomation.ViCore.Cmdlets.Commands.Lcm.ImportLcmClusterDesiredState Using vSphere Client, the image configuration can be successfully changed however.
Environment
VMware vCenter Server 8.0.x
Cause
This can happen if the cluster has a draft software configuration attached that was not properly finalized.
When this happens the cluster will be stuck in draft status and does not allow any further software configuration updates via API/SDK (including PowerCLI) call until the draft has been either committed or was removed.
Resolution
To solve the issue, you need to remove any existing draft from the cluster.
- Start with opening a new SSH connection to the vCenter Server Appliance (VCSA) and login with the root account.
- Next run the following database query to find the cluster ID (replace <clustername> with the actual display name of the cluster:
# /opt/vmware/vpostgres/current/bin/psql -d VCDB -U postgres -c "select id, name from vpx_entity where type_id = 3 and name like '%<cluster_name>%';"
Example:
root@vcsa [ /opt/vmware/vpostgres ]# /opt/vmware/vpostgres/current/bin/psql -d VCDB -U postgres -c "select id, name from vpx_entity where type_id = 3 and name like '%<clustername>%';" id | name----+------------ 8 | <clustername>(1 row)
The "id" value above is only part of the actual cluster ID. The actual one will be "domain-c<id>" - for the example it is "domain-c8".
- Now use the vSphere API to create a new API session (replace <password> against the correct password of the [email protected] account):
# curl -k -X POST "https://<vcsa_fqdn>/api/session" -u '[email protected]:<password>'
The output for this command will be a new session ID, for example:
root@vcsa [ ~ ]# curl -k -X POST "https://vcsa.example.domain/api/session" -u '[email protected]:###' "cdb21b2430598d612ba569d160eb7a54"
- Using this session ID, query the cluster for existing drafts:
# curl -k -X POST 'https://<vcsa_fqdn>/api/esx/settings/clusters/<cluster_id>/software/drafts' -H 'vmware-api-session-id: <session_id>'
For example:
# curl -k -X POST 'https://vcsa.example.domain/api/esx/settings/clusters/domain-c8/software/drafts' -H 'vmware-api-session-id: cdb21b2430598d612ba569d160eb7a54'"24"
- Use this ID to delete the current draft:
# curl -k -X DELETE 'https://<vcsa_fqdn>/api/esx/settings/clusters/<cluster_id>/software/drafts/<draft_id>' -H 'vmware-api-session-id: <session_id>'
Going with the example above:
# curl -k -X DELETE 'https://vcsa.example.domain/api/esx/settings/clusters/domain-c8/software/drafts/24' -H 'vmware-api-session-id: cdb21b2430598d612ba569d160eb7a54'
- Verify that there are no drafts left for the cluster:
# curl -k -X POST 'https://<vcsa_fqdn>/api/esx/settings/clusters/<cluster_id>/software/drafts' -H 'vmware-api-session-id: <session_id>'
Once this has been confirmed, run the Import-LcmClusterDesiredState cmdlet again in order to update the desired software configuration of the cluster.
Additional Information
Feedback
