7544717b96
Previously, when reaping child processes, if there were no remaining reapable children, the parent strand would only nap 1, which puts unnecessary load on respirate unless at least one child strand exits in the next second. Change this approach by having the exiting child strands, after they release the lease, schedule their parent immediately if the parent has no non-exited child strands. When doing this, you need to be careful to make sure there are not race conditions that would delay the scheduling of the parent. There are two potential situations you need to handle: 1. Multiple children exiting at the same time 2. Parent currently running while child is exiting By waiting until after the child strand leases are released, you still have a race condition with 1, but the race condition is that multiple child strands exiting concurrently could both reschedule the parent strand. However, that isn't a problem. You want to avoid the case where neither child strand schedules the parent, which rescheduling after releasing the lease should do. To handle 2, inside reap use Model#lock! to lock the parent strand. This will make exiting child strands block if they UPDATE the parent strand with a new schedule, until the parent strand's transaction commits. However, it's possible that a child strand already UPDATED the parent. To handle this situation, before calling lock!, store the cached schedule value in a local variable. lock! implicitly does a reload, so compare the schedule value after reload. If the schedule has changed, likely a child scheduled the parent for immediate execution, so nap 0 in that case. Just in case there are unforeseen race conditions that are not handled, only nap for 120 seconds if there are active children. Worst case scenario, this results in a 2 minute delay for running the parent. However, this can potentially result in 120x less load from parent strands polling children.
551 lines
27 KiB
Ruby
551 lines
27 KiB
Ruby
# frozen_string_literal: true
|
|
|
|
require "spec_helper"
|
|
require_relative "../../../prog/ai/inference_endpoint_replica_nexus"
|
|
|
|
RSpec.describe Prog::Ai::InferenceEndpointReplicaNexus do
|
|
subject(:nx) { described_class.new(st) }
|
|
|
|
let(:st) { Strand.create(prog: "Prog::Ai::InferenceEndpointReplicaNexus", label: "start") }
|
|
|
|
let(:project) { Project.create(name: "test") }
|
|
let(:private_subnet) { PrivateSubnet.create(project_id: project.id, name: "test", location_id: Location::HETZNER_HEL1_ID, net6: "fe80::/64", net4: "192.168.0.0/24") }
|
|
let(:load_balancer) { Prog::Vnet::LoadBalancerNexus.assemble(private_subnet.id, name: "test", src_port: 443, dst_port: 8443).subject }
|
|
|
|
let(:inference_endpoint) {
|
|
InferenceEndpoint.create(
|
|
model_name: "test-model",
|
|
replica_count: 2,
|
|
is_public: true,
|
|
location: Location[name: "hetzner-ai"],
|
|
name: "ie-name",
|
|
engine: "vllm",
|
|
engine_params: "--some-params",
|
|
external_config: {"some" => "config"},
|
|
boot_image: "image",
|
|
max_requests: 500,
|
|
max_project_rps: 100,
|
|
max_project_tps: 10000,
|
|
vm_size: "size",
|
|
storage_volumes: [],
|
|
project_id: project.id,
|
|
load_balancer_id: load_balancer.id,
|
|
private_subnet_id: private_subnet.id
|
|
)
|
|
}
|
|
|
|
let(:vm) {
|
|
vm_host = create_vm_host
|
|
vm = Prog::Vm::Nexus.assemble("k y", project.id, name: "name", private_subnet_id: private_subnet.id).subject
|
|
vm.update(vm_host_id: vm_host.id)
|
|
vm
|
|
}
|
|
|
|
let(:replica) {
|
|
InferenceEndpointReplica.create(
|
|
inference_endpoint_id: inference_endpoint.id,
|
|
vm_id: vm.id
|
|
)
|
|
}
|
|
|
|
let(:sshable) { instance_double(Sshable, host: "3.4.5.6") }
|
|
|
|
before do
|
|
allow(nx).to receive_messages(vm: vm, inference_endpoint: inference_endpoint, inference_endpoint_replica: replica)
|
|
allow(vm).to receive(:sshable).and_return(sshable)
|
|
load_balancer.add_vm(vm)
|
|
end
|
|
|
|
describe ".assemble" do
|
|
it "creates replica and vm with sshable" do
|
|
user_project = Project.create_with_id(name: "default")
|
|
ie_project = Project.create_with_id(name: "default")
|
|
Firewall.create_with_id(name: "inference-endpoint-firewall", location_id: Location::HETZNER_FSN1_ID, project_id: ie_project.id)
|
|
|
|
expect(Config).to receive(:inference_endpoint_service_project_id).and_return(ie_project.id).at_least(:once)
|
|
st_ie = Prog::Ai::InferenceEndpointNexus.assemble_with_model(
|
|
project_id: user_project.id,
|
|
location_id: Location::HETZNER_FSN1_ID,
|
|
name: "ie1",
|
|
model_id: "8b0b55b3-fb99-415f-8441-3abef2c2a200"
|
|
)
|
|
ie = st_ie.subject
|
|
st = described_class.assemble(ie.id)
|
|
replica = InferenceEndpointReplica[st.id]
|
|
expect(replica).not_to be_nil
|
|
expect(replica.vm).not_to be_nil
|
|
expect(replica.vm.sshable).not_to be_nil
|
|
expect(ie.replicas).to include(replica)
|
|
expect(ie.load_balancer.vms).to include(replica.vm)
|
|
expect(replica.vm.private_subnets).to include(ie.private_subnet)
|
|
expect(replica.vm.boot_image).to eq(ie.boot_image)
|
|
end
|
|
end
|
|
|
|
describe "#before_run" do
|
|
it "hops to destroy when needed" do
|
|
expect(nx).to receive(:when_destroy_set?).and_yield
|
|
expect { nx.before_run }.to hop("destroy")
|
|
end
|
|
|
|
it "does not hop to destroy if already in the destroy state" do
|
|
expect(nx).to receive(:when_destroy_set?).and_yield
|
|
expect(nx.strand).to receive(:label).and_return("destroy")
|
|
expect { nx.before_run }.not_to hop("destroy")
|
|
end
|
|
|
|
it "pops additional operations from stack" do
|
|
expect(nx).to receive(:when_destroy_set?).and_yield
|
|
expect(nx.strand).to receive(:label).and_return("destroy")
|
|
expect(nx.strand.stack).to receive(:count).and_return(2)
|
|
expect { nx.before_run }.to exit({"msg" => "operation is cancelled due to the destruction of the inference endpoint replica"})
|
|
end
|
|
end
|
|
|
|
describe "#start" do
|
|
it "naps if vm not ready" do
|
|
vm.strand.update(label: "prep")
|
|
expect { nx.start }.to nap(5)
|
|
end
|
|
|
|
it "update sshable host and hops" do
|
|
vm.strand.update(label: "wait")
|
|
expect { nx.start }.to hop("bootstrap_rhizome")
|
|
end
|
|
end
|
|
|
|
describe "#bootstrap_rhizome" do
|
|
it "buds a bootstrap rhizome process" do
|
|
expect(nx).to receive(:bud).with(Prog::BootstrapRhizome, {"target_folder" => "inference_endpoint", "subject_id" => replica.vm.id, "user" => "ubi"})
|
|
expect { nx.bootstrap_rhizome }.to hop("wait_bootstrap_rhizome")
|
|
end
|
|
end
|
|
|
|
describe "#wait_bootstrap_rhizome" do
|
|
it "hops to setup if there are no sub-programs running" do
|
|
expect { nx.wait_bootstrap_rhizome }.to hop("download_lb_cert")
|
|
end
|
|
|
|
it "donates if there are sub-programs running" do
|
|
Strand.create(parent_id: st.id, prog: "BootstrapRhizome", label: "start", stack: [{}], lease: Time.now + 10)
|
|
expect { nx.wait_bootstrap_rhizome }.to nap(120)
|
|
end
|
|
end
|
|
|
|
describe "#download_lb_cert" do
|
|
it "downloads lb cert and hops to setup_external" do
|
|
expect(sshable).to receive(:cmd).with("sudo inference_endpoint/bin/download-lb-cert")
|
|
expect { nx.download_lb_cert }.to hop("setup_external")
|
|
end
|
|
end
|
|
|
|
describe "#setup_external" do
|
|
it "hops to setup for vllm" do
|
|
expect { nx.setup_external }.to hop("setup")
|
|
end
|
|
|
|
it "creates a pod for runpod" do
|
|
expect(inference_endpoint).to receive(:engine).and_return "runpod"
|
|
expect(Config).to receive(:operator_ssh_public_keys).and_return "operator ssh key"
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"query Pods { myself { pods { id name runtime { ports { ip isIpPublic privatePort publicPort type } } } } }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {data: {myself: {pods: []}}}.to_json, headers: {})
|
|
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"mutation {\\n podFindAndDeployOnDemand(\\n input: {\\n cloudType: ALL\\n dataCenterId: \\\"\\\"\\n gpuCount: \\n gpuTypeId: \\\"\\\"\\n containerDiskInGb: \\n minVcpuCount: \\n minMemoryInGb: \\n imageName: \\\"\\\"\\n name: \\\"#{replica.ubid}\\\"\\n volumeInGb: 0\\n ports: \\\"22/tcp\\\"\\n env: [\\n { key: \\\"HF_TOKEN\\\", value: \\\"\\\" },\\n { key: \\\"HF_HUB_ENABLE_HF_TRANSFER\\\", value: \\\"1\\\"},\\n { key: \\\"MODEL_PATH\\\", value: \\\"/model\\\"},\\n { key: \\\"MODEL_NAME_HF\\\", value: \\\"\\\"},\\n { key: \\\"VLLM_PARAMS\\\", value: \\\"--served-model-name test-model --disable-log-requests --host 127.0.0.1 --some-params\\\"},\\n { key: \\\"SSH_KEYS\\\", value: \\\"vm ssh key\\\\noperator ssh key\\\" }\\n ]\\n }\\n ) {\\n id\\n imageName\\n env\\n machineId\\n machine {\\n podHostId\\n }\\n }\\n}\\n\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {"data" => {"podFindAndDeployOnDemand" => {"id" => "thepodid"}}}.to_json, headers: {})
|
|
expect(replica).to receive(:update).with(external_state: {"pod_id" => "thepodid"})
|
|
expect(sshable).to receive(:cmd).and_return("vm ssh key\n")
|
|
expect { nx.setup_external }.to nap(10)
|
|
end
|
|
|
|
it "does not create a pod if one already exists" do
|
|
expect(inference_endpoint).to receive(:engine).and_return "runpod"
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"query Pods { myself { pods { id name runtime { ports { ip isIpPublic privatePort publicPort type } } } } }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {"data" => {"myself" => {"pods" => [{"name" => replica.ubid.to_s, "id" => "thepodid"}]}}}.to_json, headers: {})
|
|
|
|
expect(replica).to receive(:update).with(external_state: {"pod_id" => "thepodid"})
|
|
expect { nx.setup_external }.to nap(10)
|
|
end
|
|
|
|
it "hops to setup for runpod" do
|
|
expect(inference_endpoint).to receive(:engine).and_return "runpod"
|
|
allow(replica).to receive(:external_state).and_return({"pod_id" => "thepodid"})
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"query Pod { pod(input: {podId: \\\"thepodid\\\"}) { id name runtime { ports { ip isIpPublic privatePort publicPort type } } } }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {"data" => {"pod" => {"id" => "thepodid", "runtime" => {"ports" => [{"type" => "tcp", "isIpPublic" => true, "publicPort" => 1234, "ip" => "1.2.3.4"}]}}}}.to_json, headers: {})
|
|
|
|
expect(replica).to receive(:update).with(external_state: {ip: "1.2.3.4", pod_id: "thepodid", port: 1234})
|
|
expect { nx.setup_external }.to hop("setup")
|
|
end
|
|
|
|
it "waits until runpod ip and port are available" do
|
|
expect(inference_endpoint).to receive(:engine).and_return "runpod"
|
|
allow(replica).to receive(:external_state).and_return({"pod_id" => "thepodid"})
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"query Pod { pod(input: {podId: \\\"thepodid\\\"}) { id name runtime { ports { ip isIpPublic privatePort publicPort type } } } }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {"data" => {"pod" => {"id" => "thepodid", "runtime" => {"ports" => [{"type" => "tcp", "isIpPublic" => false, "publicPort" => 1234, "ip" => "1.2.3.4"}]}}}}.to_json, headers: {})
|
|
expect { nx.setup_external }.to nap(10)
|
|
end
|
|
|
|
it "raises an error if runpod pod id is unexpected" do
|
|
expect(inference_endpoint).to receive(:engine).and_return "runpod"
|
|
allow(replica).to receive(:external_state).and_return({"pod_id" => "thepodid"})
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"query Pod { pod(input: {podId: \\\"thepodid\\\"}) { id name runtime { ports { ip isIpPublic privatePort publicPort type } } } }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {"data" => {"pod" => {"id" => "anotherpodid", "runtime" => {"ports" => [{"type" => "tcp", "isIpPublic" => false, "publicPort" => 1234, "ip" => "1.2.3.4"}]}}}}.to_json, headers: {})
|
|
expect { nx.setup_external }.to raise_error("BUG: unexpected pod id")
|
|
end
|
|
|
|
it "raises an error if pod cannot be found" do
|
|
expect(inference_endpoint).to receive(:engine).and_return "runpod"
|
|
allow(replica).to receive(:external_state).and_return({"pod_id" => "thepodid"})
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"query Pod { pod(input: {podId: \\\"thepodid\\\"}) { id name runtime { ports { ip isIpPublic privatePort publicPort type } } } }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: {"data" => {}}.to_json, headers: {})
|
|
expect { nx.setup_external }.to raise_error("BUG: pod not found")
|
|
end
|
|
end
|
|
|
|
describe "#setup" do
|
|
it "triggers setup if setup command is not sent yet or failed" do
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer 'sudo inference_endpoint/bin/setup-replica' setup", {stdin: "{\"engine_start_cmd\":\"/opt/miniconda/envs/vllm/bin/vllm serve /ie/models/model --served-model-name llama --disable-log-requests --host 127.0.0.1 --some-params\",\"replica_ubid\":\"#{replica.ubid}\",\"ssl_crt_path\":\"/ie/workdir/ssl/ubi_cert.pem\",\"ssl_key_path\":\"/ie/workdir/ssl/ubi_key.pem\",\"gateway_port\":8443,\"max_requests\":500}"}).twice
|
|
expect(inference_endpoint).to receive(:gpu_count).and_return(1).twice
|
|
expect(inference_endpoint).to receive(:engine).and_return("vllm").twice
|
|
expect(inference_endpoint).to receive(:engine_params).and_return("--some-params").twice
|
|
expect(inference_endpoint).to receive(:model_name).and_return("llama").twice
|
|
|
|
# NotStarted
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("NotStarted")
|
|
expect { nx.setup }.to nap(5)
|
|
|
|
# Failed
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("Failed")
|
|
expect { nx.setup }.to nap(5)
|
|
end
|
|
|
|
it "triggers setup for vllm with cpu if setup command is not sent yet or failed" do
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer 'sudo inference_endpoint/bin/setup-replica' setup", {stdin: "{\"engine_start_cmd\":\"/opt/miniconda/envs/vllm-cpu/bin/vllm serve /ie/models/model --served-model-name llama --disable-log-requests --host 127.0.0.1 --some-params\",\"replica_ubid\":\"#{replica.ubid}\",\"ssl_crt_path\":\"/ie/workdir/ssl/ubi_cert.pem\",\"ssl_key_path\":\"/ie/workdir/ssl/ubi_key.pem\",\"gateway_port\":8443,\"max_requests\":500}"})
|
|
expect(inference_endpoint).to receive(:gpu_count).and_return(0)
|
|
expect(inference_endpoint).to receive(:engine).and_return("vllm")
|
|
expect(inference_endpoint).to receive(:engine_params).and_return("--some-params")
|
|
expect(inference_endpoint).to receive(:model_name).and_return("llama")
|
|
|
|
# NotStarted
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("NotStarted")
|
|
expect { nx.setup }.to nap(5)
|
|
end
|
|
|
|
it "triggers setup for runpod if setup command is not sent yet or failed" do
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer 'sudo inference_endpoint/bin/setup-replica' setup", {stdin: "{\"engine_start_cmd\":\"ssh -N -L 8000:localhost:8000 root@ -p -i /ie/workdir/.ssh/runpod -o UserKnownHostsFile=/ie/workdir/.ssh/known_hosts -o StrictHostKeyChecking=accept-new\",\"replica_ubid\":\"#{replica.ubid}\",\"ssl_crt_path\":\"/ie/workdir/ssl/ubi_cert.pem\",\"ssl_key_path\":\"/ie/workdir/ssl/ubi_key.pem\",\"gateway_port\":8443,\"max_requests\":500}"}).twice
|
|
expect(inference_endpoint).to receive(:engine).and_return("runpod").twice
|
|
|
|
# NotStarted
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("NotStarted")
|
|
expect { nx.setup }.to nap(5)
|
|
|
|
# Failed
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("Failed")
|
|
expect { nx.setup }.to nap(5)
|
|
end
|
|
|
|
it "hops to wait_endpoint_up if setup command has succeeded" do
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("Succeeded")
|
|
expect { nx.setup }.to hop("wait_endpoint_up")
|
|
end
|
|
|
|
it "naps if script return unknown status" do
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("Unknown")
|
|
expect { nx.setup }.to nap(5)
|
|
end
|
|
|
|
it "fails if inference engine is unsupported" do
|
|
expect(sshable).to receive(:cmd).with("common/bin/daemonizer --check setup").and_return("NotStarted")
|
|
expect(inference_endpoint).to receive(:engine).and_return("unsupported engine")
|
|
expect { nx.setup }.to raise_error("BUG: unsupported inference engine")
|
|
end
|
|
end
|
|
|
|
describe "#wait_endpoint_up" do
|
|
it "naps if vm is not up" do
|
|
LoadBalancerVmPort.first.update(state: "down")
|
|
expect { nx.wait_endpoint_up }.to nap(5)
|
|
end
|
|
|
|
it "sets hops to wait when vm is in active set of load balancer" do
|
|
LoadBalancerVmPort.first.update(state: "up")
|
|
expect { nx.wait_endpoint_up }.to hop("wait")
|
|
end
|
|
end
|
|
|
|
describe "#wait" do
|
|
it "pings the inference gateway and naps" do
|
|
expect(nx).to receive(:available?).and_return(true)
|
|
expect(nx).to receive(:ping_gateway)
|
|
expect { nx.wait }.to nap(120)
|
|
end
|
|
|
|
it "hops to unavailable if the replica is not available" do
|
|
expect(nx).to receive(:available?).and_return(false)
|
|
expect { nx.wait }.to hop("unavailable")
|
|
end
|
|
end
|
|
|
|
describe "#unavailable" do
|
|
it "creates a page if replica is unavailable" do
|
|
LoadBalancerVmPort.first.update(state: "down")
|
|
expect(Prog::PageNexus).to receive(:assemble)
|
|
expect(inference_endpoint).to receive(:maintenance_set?).and_return(false)
|
|
expect { nx.unavailable }.to nap(30)
|
|
end
|
|
|
|
it "does not create a page if replica is in maintenance mode" do
|
|
LoadBalancerVmPort.first.update(state: "down")
|
|
expect(Prog::PageNexus).not_to receive(:assemble)
|
|
expect(inference_endpoint).to receive(:maintenance_set?).and_return(true)
|
|
expect { nx.unavailable }.to nap(30)
|
|
end
|
|
|
|
it "resolves the page if replica is available" do
|
|
pg = instance_double(Page)
|
|
expect(pg).to receive(:incr_resolve)
|
|
expect(nx).to receive(:available?).and_return(true)
|
|
expect(Page).to receive(:from_tag_parts).and_return(pg)
|
|
expect { nx.unavailable }.to hop("wait")
|
|
end
|
|
|
|
it "does not resolves the page if there is none" do
|
|
expect(nx).to receive(:available?).and_return(true)
|
|
expect(Page).to receive(:from_tag_parts).and_return(nil)
|
|
expect { nx.unavailable }.to hop("wait")
|
|
end
|
|
end
|
|
|
|
describe "#destroy" do
|
|
it "deletes resources and exits" do
|
|
lb = instance_double(LoadBalancer)
|
|
expect(inference_endpoint).to receive(:load_balancer).and_return(lb).twice
|
|
expect(lb).to receive(:evacuate_vm).with(vm)
|
|
expect(lb).to receive(:remove_vm).with(vm)
|
|
|
|
expect(vm).to receive(:incr_destroy)
|
|
expect(replica).to receive(:destroy)
|
|
|
|
expect { nx.destroy }.to exit({"msg" => "inference endpoint replica is deleted"})
|
|
end
|
|
|
|
it "deletes runpod pod if there is one" do
|
|
lb = instance_double(LoadBalancer)
|
|
expect(inference_endpoint).to receive(:load_balancer).and_return(lb).twice
|
|
expect(lb).to receive(:evacuate_vm).with(vm)
|
|
expect(lb).to receive(:remove_vm).with(vm)
|
|
expect(replica).to receive(:external_state).and_return({"pod_id" => "thepodid"})
|
|
|
|
stub_request(:post, "https://api.runpod.io/graphql")
|
|
.with(
|
|
body: "{\"query\":\"mutation { podTerminate(input: {podId: \\\"thepodid\\\"}) }\"}",
|
|
headers: {
|
|
"Accept-Encoding" => "deflate, gzip",
|
|
"Authorization" => "Bearer ",
|
|
"Content-Type" => "application/json",
|
|
"Host" => "api.runpod.io"
|
|
}
|
|
)
|
|
.to_return(status: 200, body: "", headers: {})
|
|
|
|
expect(replica).to receive(:update).with(external_state: "{}")
|
|
expect(vm).to receive(:incr_destroy)
|
|
expect(replica).to receive(:destroy)
|
|
|
|
expect { nx.destroy }.to exit({"msg" => "inference endpoint replica is deleted"})
|
|
end
|
|
end
|
|
|
|
describe "#ping_gateway" do
|
|
let(:projects) { [Project.create_with_id(name: "p1"), Project.create_with_id(name: "p2")] }
|
|
|
|
before do
|
|
ApiKey.create_inference_api_key(projects.first)
|
|
ApiKey.create_inference_api_key(projects.last)
|
|
end
|
|
|
|
it "for private endpoints" do
|
|
expect(inference_endpoint).to receive(:project).and_return(projects.first)
|
|
expect(inference_endpoint).to receive(:is_public).and_return(false).twice
|
|
expect(inference_endpoint).to receive(:ubid).and_return("ieubid")
|
|
expect(nx).to receive(:update_billing_records).with(
|
|
JSON.parse("[{\"ubid\":\"#{replica.ubid}\",\"request_count\":1,\"prompt_token_count\":10,\"completion_token_count\":20},{\"ubid\":\"anotherubid\",\"request_count\":0,\"prompt_token_count\":0,\"completion_token_count\":0}]"),
|
|
"input", "prompt_token_count"
|
|
)
|
|
expect(nx).to receive(:update_billing_records).with(
|
|
JSON.parse("[{\"ubid\":\"#{replica.ubid}\",\"request_count\":1,\"prompt_token_count\":10,\"completion_token_count\":20},{\"ubid\":\"anotherubid\",\"request_count\":0,\"prompt_token_count\":0,\"completion_token_count\":0}]"),
|
|
"output", "completion_token_count"
|
|
)
|
|
expect(sshable).to receive(:cmd).with("sudo curl -m 10 --no-progress-meter -H \"Content-Type: application/json\" -X POST --data-binary @- --unix-socket /ie/workdir/inference-gateway.clover.sock http://localhost/control", {stdin: "{\"replica_ubid\":\"#{replica.ubid}\",\"public_endpoint\":false,\"projects\":[{\"ubid\":\"#{projects.first.ubid}\",\"api_keys\":[\"#{Digest::SHA2.hexdigest(projects.first.api_keys.first.key)}\"],\"quota_rps\":100,\"quota_tps\":10000}]}"}).and_return("{\"inference_endpoint\":\"1eqhk4b9gfq27gc5agxkq84bhr\",\"replica\":\"1rvtmbhd8cne6jpz3xxat7rsnr\",\"projects\":[{\"ubid\":\"#{replica.ubid}\",\"request_count\":1,\"prompt_token_count\":10,\"completion_token_count\":20},{\"ubid\":\"anotherubid\",\"request_count\":0,\"prompt_token_count\":0,\"completion_token_count\":0}]}")
|
|
nx.ping_gateway
|
|
end
|
|
|
|
it "for public endpoints" do
|
|
expect(inference_endpoint).to receive(:is_public).and_return(true).twice
|
|
expect(inference_endpoint).to receive(:ubid).and_return("ieubid")
|
|
|
|
expected_projects = [
|
|
{"ubid" => projects.first.ubid, "api_keys" => [Digest::SHA2.hexdigest(projects.first.api_keys.first.key)], "quota_rps" => 100, "quota_tps" => 10000},
|
|
{"ubid" => projects.last.ubid, "api_keys" => [Digest::SHA2.hexdigest(projects.last.api_keys.first.key)], "quota_rps" => 100, "quota_tps" => 10000}
|
|
].sort_by { |p| p["ubid"] }
|
|
|
|
expect(sshable).to receive(:cmd) do |command, options|
|
|
json_sent = JSON.parse(options[:stdin])
|
|
projects_sent = json_sent["projects"].sort_by { |p| p["ubid"] }
|
|
expect(projects_sent).to eq(expected_projects)
|
|
end.and_return("{\"inference_endpoint\":\"1eqhk4b9gfq27gc5agxkq84bhr\",\"replica\":\"1rvtmbhd8cne6jpz3xxat7rsnr\",\"projects\":[{\"ubid\":\"#{replica.ubid}\",\"request_count\":1,\"prompt_token_count\":10,\"completion_token_count\":20},{\"ubid\":\"anotherubid\",\"request_count\":0,\"prompt_token_count\":0,\"completion_token_count\":0}]}")
|
|
expect(nx).to receive(:update_billing_records).with(
|
|
JSON.parse("[{\"ubid\":\"#{replica.ubid}\",\"request_count\":1,\"prompt_token_count\":10,\"completion_token_count\":20},{\"ubid\":\"anotherubid\",\"request_count\":0,\"prompt_token_count\":0,\"completion_token_count\":0}]"),
|
|
"input", "prompt_token_count"
|
|
)
|
|
expect(nx).to receive(:update_billing_records).with(
|
|
JSON.parse("[{\"ubid\":\"#{replica.ubid}\",\"request_count\":1,\"prompt_token_count\":10,\"completion_token_count\":20},{\"ubid\":\"anotherubid\",\"request_count\":0,\"prompt_token_count\":0,\"completion_token_count\":0}]"),
|
|
"output", "completion_token_count"
|
|
)
|
|
|
|
nx.ping_gateway
|
|
end
|
|
end
|
|
|
|
describe "#update_billing_records" do
|
|
let(:p1) { Project.create_with_id(name: "default") }
|
|
|
|
it "updates billing records" do
|
|
expect(BillingRecord.count).to eq(0)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 10, "completion_token_count" => 20}],
|
|
"input", "prompt_token_count"
|
|
)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 10, "completion_token_count" => 20}],
|
|
"output", "completion_token_count"
|
|
)
|
|
expect(BillingRecord.count).to eq(2)
|
|
br, br2 = BillingRecord.order(:billing_rate_id).all
|
|
expect(br.project_id).to eq(p1.id)
|
|
expect(br.resource_id).to eq(inference_endpoint.id)
|
|
expect(br.billing_rate_id).to eq("ba80e171-0c24-4bf9-ac4f-36bdadb259c0")
|
|
expect(br.amount).to eq(10)
|
|
expect(br2.project_id).to eq(p1.id)
|
|
expect(br2.resource_id).to eq(inference_endpoint.id)
|
|
expect(br2.billing_rate_id).to eq("c8886006-9e15-4046-b46a-163851626f83")
|
|
expect(br2.amount).to eq(20)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 1, "completion_token_count" => 2}],
|
|
"input", "prompt_token_count"
|
|
)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 1, "completion_token_count" => 2}],
|
|
"output", "completion_token_count"
|
|
)
|
|
expect(BillingRecord.count).to eq(2)
|
|
expect(Integer(br.reload.amount)).to eq(11)
|
|
expect(Integer(br2.reload.amount)).to eq(22)
|
|
end
|
|
|
|
it "does not update for zero tokens" do
|
|
expect(BillingRecord.count).to eq(0)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 0, "prompt_token_count" => 0, "completion_token_count" => 0}],
|
|
"input", "prompt_token_count"
|
|
)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 0, "prompt_token_count" => 0, "completion_token_count" => 0}],
|
|
"output", "completion_token_count"
|
|
)
|
|
expect(BillingRecord.count).to eq(0)
|
|
end
|
|
|
|
it "does not update if price is zero" do
|
|
expect(BillingRate).to receive(:from_resource_properties).with("InferenceTokens", "#{inference_endpoint.model_name}-input", "global").and_return({"unit_price" => 0.0000000000})
|
|
expect(BillingRate).to receive(:from_resource_properties).with("InferenceTokens", "#{inference_endpoint.model_name}-output", "global").and_return({"unit_price" => 0.0000000000})
|
|
expect(BillingRecord.count).to eq(0)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 2, "completion_token_count" => 3}],
|
|
"input", "prompt_token_count"
|
|
)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 2, "completion_token_count" => 3}],
|
|
"output", "completion_token_count"
|
|
)
|
|
expect(BillingRecord.count).to eq(0)
|
|
end
|
|
|
|
it "failure in updating single record doesn't impact others" do
|
|
p2 = Project.create_with_id(name: "default")
|
|
expect(BillingRecord).to receive(:create_with_id).once.ordered.with(hash_including(project_id: p1.id)).and_raise(Sequel::DatabaseConnectionError)
|
|
expect(BillingRecord).to receive(:create_with_id).once.ordered.with(hash_including(project_id: p2.id)).and_call_original
|
|
expect(BillingRecord.count).to eq(0)
|
|
nx.update_billing_records(
|
|
[{"ubid" => p1.ubid, "request_count" => 1, "prompt_token_count" => 2, "completion_token_count" => 3}, {"ubid" => p2.ubid, "request_count" => 1, "prompt_token_count" => 2, "completion_token_count" => 3}],
|
|
"input", "prompt_token_count"
|
|
)
|
|
expect(BillingRecord.count).to eq(1)
|
|
br = BillingRecord.first
|
|
expect(br.project_id).to eq(p2.id)
|
|
end
|
|
end
|
|
end
|