Virtual machine errors

Causes and remedies for the error codes of virtual machine power operations, placement, configuration limits and snapshots.

`error.invalid_power_state`

Text in the console: The operation is not possible while the virtual machine is {power_state, select, powered_on {powered on} PoweredOn {powered on} poweredon {powered on} powered_off {powered off} PoweredOff {powered off} poweredoff {powered off} suspended {suspended} Suspended {suspended} paused {paused} Paused {paused} other {in this power state}}. {required, select, powered_off {Power off the virtual machine} powered_on {Power on the virtual machine} suspended {Suspend the virtual machine} other {Change the power state}} first, then try again.
Meaning for API clients: The operation needs another power state: power_state is the current one, required the one it needs (powered_off, powered_on or suspended).
Code: FAILED_PRECONDITION, HTTP 412

Argument	What it holds
`power_state`	the current power state, for example `poweredon` or `poweredoff`.
`required`	the power state the operation needs: `powered_off`, `powered_on` or `suspended`.

Cause. The operation is not possible in the current power state of the virtual machine. The required argument names the state it needs: powered_off for unregistering, converting to a template, customizing, destroying, moving storage and cloning; powered_on for reset and suspend. For power-on it is powered_off and for power-off powered_on, although a suspended or paused machine also qualifies for both.

Remedy.

Check the power state in the inventory: kyvernon inventory vm <id>.
Bring the virtual machine to the state in required, for example Power off or kyvernon vm power-off <id>, then repeat the operation.
When the state changes by itself, wait for the task that runs on the virtual machine to finish: kyvernon tasks list --target <id>.

`error.vm_not_placed`

Text in the console: The virtual machine is not placed on a host. {reason, select, no_connected_host {No host of its cluster is connected. Reconnect a host, then move the virtual machine to it.} all_in_maintenance {Every connected host of its cluster is in maintenance mode. End maintenance mode on a host, then move the virtual machine to it.} no_hosts {Its cluster has no hosts. Add a host to the cluster, then move the virtual machine to it.} no_compute_resource {Its resource pool does not belong to a cluster or a standalone host. Move the virtual machine to a resource pool of a cluster or host.} other {No host is assigned to it. Move the virtual machine to a host.}}
Meaning for API clients: The VM has no host. reason is no_connected_host, all_in_maintenance, no_hosts, no_compute_resource (placement finds no host today) or not_assigned (a host is available but the VM names none).
Code: FAILED_PRECONDITION, HTTP 412

Argument	What it holds
`reason`	`no_connected_host`, `all_in_maintenance`, `no_hosts`, `no_compute_resource` or `not_assigned`.

Cause. The virtual machine has no host. The reason argument says what placement finds today:

no_connected_host: the hosts of its cluster exist but none is connected.
all_in_maintenance: every connected host of its cluster is in maintenance mode.
no_hosts: its cluster has no hosts.
no_compute_resource: its resource pool does not belong to a cluster or a standalone host.
not_assigned: a host is available, but the virtual machine record names none.

Placement looks at the connection and the maintenance mode only. Free memory is checked when you power on. Power-on does not place a virtual machine that has no host. Operations that reach into the running machine, such as the console, guest operations and verification, need a host.

Remedy.

Read the reason in the message and fix it: reconnect a host with kyvernon host reconnect <id> (no_connected_host), end maintenance mode on a host with kyvernon host exit-maintenance <id> (all_in_maintenance), add a host to the cluster (no_hosts) or move the virtual machine to a resource pool of a cluster or host (no_compute_resource).
Place the virtual machine on a host: kyvernon vm relocate <id> --host <host>.
Repeat the operation. When the reason stays, pass on the name of the virtual machine and the reason to an administrator.

`error.vm_spec_unbuildable`

Text in the console: The configuration of the virtual machine cannot be built on its host. Check the hardware settings and the host's capabilities. The task details name the cause.
Meaning for API clients: The VM configuration cannot be built on the host; the message names the cause.
Code: FAILED_PRECONDITION, HTTP 412

Cause. The configuration of the virtual machine cannot be turned into a machine on the host. The message of the API response, and the details of the failed task, give the reason that the host reported: a device or setting that the host does not support. The error carries no argument for the cause.

Remedy.

Find the cause: read the message in the response, or open the details of the failed task: kyvernon tasks list --target <id>. The log of the controller holds it too.
Check the hardware settings it names: Edit settings.
Remove or change the device or setting.
When the host is the problem, choose another host: kyvernon vm relocate <id> --host <host>.

`error.vm.concurrent_modification`

Text in the console: The configuration of the virtual machine changed while you were editing it. Reload the virtual machine and make the change again.
Meaning for API clients: The VM configuration changed since the request's change_version was read and the change cannot be merged; read the VM again and repeat (HTTP 409).
Code: CONFLICT, HTTP 409

Cause. The configuration of the virtual machine changed after you read it, and your change cannot be merged with the other one. The request carries the change_version that you read, and it is not current.

Remedy.

Read the virtual machine again: reload it in the console, or run kyvernon inventory vm <id>.
Make the change again on the current configuration: kyvernon vm reconfigure <id> --config <file>.

`error.vm.exceeds_host_cpu`

Text in the console: {requested} vCPUs are more than the {limit} that host {host} allows for one virtual machine. Choose fewer vCPUs.
Meaning for API clients: More vCPUs than the host has logical CPUs (or than vm_ops.max_vcpus_per_vm).
Code: INVALID_ARGUMENT, HTTP 400

Argument	What it holds
`field`	the request field that holds the rejected value, here `config.num_cpus`.
`requested`	the number of vCPUs asked for.
`limit`	the largest number allowed.
`host`	the name of the host.

Cause. The number of vCPUs is larger than the number of logical CPUs of the host, or larger than the limit max_vcpus_per_vm in the [vm_ops] section of the controller configuration. When that setting is 0, the logical CPUs of the host are the limit.

Remedy.

Choose fewer vCPUs: Edit settings.
Or place the virtual machine on a host with more logical CPUs.
An administrator can raise max_vcpus_per_vm. Pass on the message id and the number that you need.

`error.vm.exceeds_host_memory`

Text in the console: {requested} MiB of memory are more than the {limit} MiB that host {host} can give one virtual machine. Choose less memory.
Meaning for API clients: More memory (MiB) than the host has after vm_ops.host_memory_reserve_mib.
Code: INVALID_ARGUMENT, HTTP 400

Argument	What it holds
`field`	the request field that holds the rejected value, here `config.memory_mib`.
`requested`	the memory in MiB asked for.
`limit`	the largest amount in MiB.
`host`	the name of the host.

Cause. The memory is larger than what the host can give one virtual machine: the memory of the host minus the reserve that the host keeps for itself. The reserve is host_memory_reserve_mib in the [vm_ops] section of the controller configuration (2048 MiB by default).

Remedy.

Choose less memory: Edit settings.
Or place the virtual machine on a host with more memory.
An administrator can lower host_memory_reserve_mib. Pass on the message id and the amount that you need.

`error.vm.too_many_devices`

Text in the console: A virtual machine can have at most {limit} {device, select, cdroms {CD-ROM drives} other {disks and network adapters together}}; this configuration has {count}. Remove devices first.
Meaning for API clients: More disks and network adapters together (device = disks_and_nics), or CD-ROM drives (cdroms), than a virtual machine has slots.
Code: INVALID_ARGUMENT, HTTP 400

Argument	What it holds
`device`	`disks_and_nics` or `cdroms`.
`count`	the number of devices in the configuration.
`limit`	the largest number allowed.

Cause. The configuration has more devices than a virtual machine has slots for. Disks and network adapters together share one limit, and CD-ROM drives have their own. A configuration that is over the limit already can still change, as long as it does not grow.

Remedy.

Remove devices that are not needed: Edit settings.
Repeat the change.

`error.no_current_snapshot`

Text in the console: The virtual machine has no current snapshot to revert to. Take a snapshot first.
Meaning for API clients: The VM has no current snapshot to revert to.
Code: NOT_FOUND, HTTP 404

Cause. Reverting to the current snapshot was asked for, but the virtual machine has no current snapshot.

Remedy.

List the snapshots and revert to a named one: kyvernon vm snapshot revert <id> <key>.
To be able to revert in future, take a snapshot first: kyvernon vm snapshot create <id> --name <name>. In the console, use Snapshots.

`error.template_relocate`

Text in the console: Templates cannot be relocated
Meaning for API clients: Templates cannot be relocated.
Code: FAILED_PRECONDITION, HTTP 412

Cause. The template was to be moved to another host or datastore. Templates cannot be relocated.

Remedy.

Convert the template to a virtual machine, relocate it, and convert it back: kyvernon vm mark-as-vm <id>, kyvernon vm relocate <id> --datastore <datastore>, kyvernon vm mark-as-template <id>.
Or clone the template to the destination: kyvernon vm clone <id> --name <name> --datastore <datastore> --template.