Source
pub async fn reserve_resource( &self, component: String, name: String, action: String, bytes_requested: i64, owner_id: i64, ) -> Result<ReserveResourceResponse, KineticaError>
Request to reserve a resource (e.g. memory) on a gpudb node
Request for a component reserve a resource on a gpudb node. Currently only memory is supported.
component- Component requesting a resource reservation e.g. graph0. This may be a new reservation or change from a previous reservation.name- Identifier for the resource.action- Modification operation to be applied.bytes_requested- Absolute number of bytes requested for the reservation. Not used for evicted/deleted statuses.owner_id- Owner identifier. Only needed for new resources or change owner.
Source
pub async fn reserve_resource_with_opts( &self, component: String, name: String, action: String, bytes_requested: i64, owner_id: i64, options: ReserveResourceRequestOptions, ) -> Result<ReserveResourceResponse, KineticaError>
Request to reserve a resource (e.g. memory) on a gpudb node
Request for a component reserve a resource on a gpudb node. Currently only memory is supported.
component- Component requesting a resource reservation e.g. graph0. This may be a new reservation or change from a previous reservation.name- Identifier for the resource.action- Modification operation to be applied.bytes_requested- Absolute number of bytes requested for the reservation. Not used for evicted/deleted statuses.owner_id- Owner identifier. Only needed for new resources or change owner.options- The optional parameters; any left unset take the server’s defaults.
Source
pub async fn reserve_resource_raw( &self, request: &ReserveResourceRequest, ) -> Result<ReserveResourceResponse, KineticaError>
Sends a pre-built ReserveResourceRequest to the /reserve/resource endpoint.
Source
pub async fn reserve_resource_from_url( &self, url: &str, request: &ReserveResourceRequest, ) -> Result<ReserveResourceResponse, KineticaError>
Sends a pre-built ReserveResourceRequest to a specific worker URL, bypassing HA head-node selection. Used for multi-head access.