This is an automated email from the ASF dual-hosted git repository.
terrymanu pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/shardingsphere.git
The following commit(s) were added to refs/heads/master by this push:
new 242c11f73af Remove pseudo MCP LLM server runtime configuration (#38872)
242c11f73af is described below
commit 242c11f73af82310bc9f85d71045dff7c7fe46eb
Author: Liang Zhang <[email protected]>
AuthorDate: Thu Jun 18 23:48:42 2026 +0800
Remove pseudo MCP LLM server runtime configuration (#38872)
Remove the configurable mcp.llm.server-runtime property because the LLM E2E
Docker runtime is fixed to llama.cpp.
Also drop the related Docker build argument and workflow input, keep
llama.cpp
as explicit runtime evidence, and update affected tests accordingly.
---
.github/workflows/e2e-mcp.yml | 2 --
.../test/e2e/mcp/llm/config/LLME2EConfiguration.java | 11 +++--------
.../test/e2e/mcp/llm/config/LLME2EConfigurationTest.java | 10 +---------
.../mcp/llm/conversation/client/LLMChatModelClientTest.java | 2 +-
.../test/e2e/mcp/llm/fixture/LLMRuntimeSupport.java | 3 +--
.../test/e2e/mcp/llm/fixture/LLMRuntimeSupportTest.java | 4 ++--
.../mcp/llm/suite/usability/LLMUsabilitySuiteRunnerTest.java | 2 +-
test/e2e/mcp/src/test/resources/docker/llm-runtime/Dockerfile | 3 +--
.../mcp/src/test/resources/docker/llm-runtime/build-local.sh | 3 ---
test/e2e/mcp/src/test/resources/env/e2e-env.properties | 1 -
10 files changed, 10 insertions(+), 31 deletions(-)
diff --git a/.github/workflows/e2e-mcp.yml b/.github/workflows/e2e-mcp.yml
index b678ba4c3fb..25ab6760013 100644
--- a/.github/workflows/e2e-mcp.yml
+++ b/.github/workflows/e2e-mcp.yml
@@ -108,7 +108,6 @@ jobs:
echo "target_platform=${target_platform}"
echo "server_image=$(load_property 'mcp.llm.server-image')"
echo "base_server_image=${base_server_image}"
- echo "server_runtime=$(load_property 'mcp.llm.server-runtime')"
echo "model_repository=$(load_property 'mcp.llm.model-repository')"
echo "model_quantization=$(load_property
'mcp.llm.model-quantization')"
echo "model_reference=$(load_property 'mcp.llm.model')"
@@ -125,7 +124,6 @@ jobs:
-f test/e2e/mcp/src/test/resources/docker/llm-runtime/Dockerfile \
-t "${{ steps.mcp-e2e-properties.outputs.server_image }}" \
--build-arg "BASE_IMAGE=${{
steps.mcp-e2e-properties.outputs.base_server_image }}" \
- --build-arg "SERVER_RUNTIME=${{
steps.mcp-e2e-properties.outputs.server_runtime }}" \
--build-arg "MODEL_REPOSITORY=${{
steps.mcp-e2e-properties.outputs.model_repository }}" \
--build-arg "MODEL_QUANTIZATION=${{
steps.mcp-e2e-properties.outputs.model_quantization }}" \
--build-arg "MODEL_REFERENCE=${{
steps.mcp-e2e-properties.outputs.model_reference }}" \
diff --git
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfiguration.java
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfiguration.java
index a14afed0745..e5e029d31db 100644
---
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfiguration.java
+++
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfiguration.java
@@ -48,8 +48,6 @@ public final class LLME2EConfiguration {
private static final String DEFAULT_API_KEY = "mcp-llm-score";
- private static final String DEFAULT_SERVER_RUNTIME = "llama.cpp";
-
private static final String DEFAULT_SERVER_IMAGE =
"apache/shardingsphere-mcp-llm-runtime:local";
private final String baseUrl;
@@ -72,8 +70,6 @@ public final class LLME2EConfiguration {
private final RuntimeMode runtimeMode;
- private final String serverRuntime;
-
private final String serverImage;
private final String baseServerImage;
@@ -102,7 +98,6 @@ public final class LLME2EConfiguration {
Paths.get(readString(props, "mcp.llm.artifact-root",
"target/llm-e2e")),
readString(props, "mcp.llm.run-id", createDefaultRunId()),
runtimeMode,
- readString(props, "mcp.llm.server-runtime",
DEFAULT_SERVER_RUNTIME),
readString(props, "mcp.llm.server-image",
DEFAULT_SERVER_IMAGE),
readString(props, "mcp.llm.base-server-image", ""),
readString(props, "mcp.llm.base-server-image-digest", ""),
@@ -130,7 +125,7 @@ public final class LLME2EConfiguration {
*/
public LLME2EConfiguration withBaseUrl(final String baseUrl) {
return new LLME2EConfiguration(normalizeBaseUrl(baseUrl),
modelProvider, modelName, apiKey, readyTimeoutSeconds, requestTimeoutSeconds,
maxTurns, artifactRoot, runId,
- runtimeMode, serverRuntime, serverImage, baseServerImage,
baseServerImageDigest, modelMetadata);
+ runtimeMode, serverImage, baseServerImage,
baseServerImageDigest, modelMetadata);
}
/**
@@ -142,7 +137,7 @@ public final class LLME2EConfiguration {
*/
public LLME2EConfiguration withModelEndpoint(final String baseUrl, final
String apiKey) {
return new LLME2EConfiguration(normalizeBaseUrl(baseUrl),
modelProvider, modelName, apiKey, readyTimeoutSeconds, requestTimeoutSeconds,
maxTurns, artifactRoot, runId,
- runtimeMode, serverRuntime, serverImage, baseServerImage,
baseServerImageDigest, modelMetadata);
+ runtimeMode, serverImage, baseServerImage,
baseServerImageDigest, modelMetadata);
}
/**
@@ -154,7 +149,7 @@ public final class LLME2EConfiguration {
*/
public LLME2EConfiguration withReadinessTimeouts(final int
readyTimeoutSeconds, final int requestTimeoutSeconds) {
return new LLME2EConfiguration(baseUrl, modelProvider, modelName,
apiKey, readyTimeoutSeconds, requestTimeoutSeconds, maxTurns, artifactRoot,
runId, runtimeMode,
- serverRuntime, serverImage, baseServerImage,
baseServerImageDigest, modelMetadata);
+ serverImage, baseServerImage, baseServerImageDigest,
modelMetadata);
}
/**
diff --git
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfigurationTest.java
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfigurationTest.java
index ebec14b662f..78dfeff7ff7 100644
---
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfigurationTest.java
+++
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/config/LLME2EConfigurationTest.java
@@ -53,8 +53,6 @@ class LLME2EConfigurationTest {
private String originalBaseServerImageDigest;
- private String originalServerRuntime;
-
private String originalModelRepository;
private String originalModelFileName;
@@ -74,7 +72,6 @@ class LLME2EConfigurationTest {
originalServerImage = System.getProperty("mcp.llm.server-image");
originalBaseServerImage =
System.getProperty("mcp.llm.base-server-image");
originalBaseServerImageDigest =
System.getProperty("mcp.llm.base-server-image-digest");
- originalServerRuntime = System.getProperty("mcp.llm.server-runtime");
originalModelRepository =
System.getProperty("mcp.llm.model-repository");
originalModelFileName = System.getProperty("mcp.llm.model-file-name");
originalModelQuantization =
System.getProperty("mcp.llm.model-quantization");
@@ -86,7 +83,6 @@ class LLME2EConfigurationTest {
System.clearProperty("mcp.llm.server-image");
System.clearProperty("mcp.llm.base-server-image");
System.clearProperty("mcp.llm.base-server-image-digest");
- System.clearProperty("mcp.llm.server-runtime");
System.clearProperty("mcp.llm.model-repository");
System.clearProperty("mcp.llm.model-file-name");
System.clearProperty("mcp.llm.model-quantization");
@@ -103,7 +99,6 @@ class LLME2EConfigurationTest {
restoreProperty("mcp.llm.server-image", originalServerImage);
restoreProperty("mcp.llm.base-server-image", originalBaseServerImage);
restoreProperty("mcp.llm.base-server-image-digest",
originalBaseServerImageDigest);
- restoreProperty("mcp.llm.server-runtime", originalServerRuntime);
restoreProperty("mcp.llm.model-repository", originalModelRepository);
restoreProperty("mcp.llm.model-file-name", originalModelFileName);
restoreProperty("mcp.llm.model-quantization",
originalModelQuantization);
@@ -121,7 +116,6 @@ class LLME2EConfigurationTest {
String expectedModelReference =
expectedProps.getProperty("mcp.llm.model");
assertThat(actual.getModelName(), is(expectedModelReference));
assertThat(actual.getApiKey(), is("mcp-llm-score"));
- assertThat(actual.getServerRuntime(),
is(expectedProps.getProperty("mcp.llm.server-runtime")));
assertThat(actual.getServerImage(),
is("apache/shardingsphere-mcp-llm-runtime:local"));
assertThat(actual.getBaseServerImage(),
is("ghcr.io/ggml-org/llama.cpp:server-b9191"));
assertThat(actual.getBaseServerImageDigest(),
is(expectedProps.getProperty("mcp.llm.base-server-image-digest")));
@@ -167,14 +161,12 @@ class LLME2EConfigurationTest {
System.setProperty("mcp.llm.server-image",
"test/mcp-llm-runtime:test");
System.setProperty("mcp.llm.base-server-image", "test/llama.cpp:test");
System.setProperty("mcp.llm.base-server-image-digest",
"test-base-server-image-digest");
- System.setProperty("mcp.llm.server-runtime", "test-runtime");
System.setProperty("mcp.llm.model-repository",
"ggml-org/Qwen3-1.7B-GGUF");
System.setProperty("mcp.llm.model-file-name",
"Qwen3-1.7B-Q4_K_M.gguf");
System.setProperty("mcp.llm.model-quantization", "Q4_K_M");
System.setProperty("mcp.llm.model-revision",
"daeb8e2d528a760970442092f6bf1e55c3b659eb");
System.setProperty("mcp.llm.model-sha256", "configured-model-sha256");
LLME2EConfiguration actual = LLME2EConfiguration.load();
- assertThat(actual.getServerRuntime(), is("test-runtime"));
assertThat(actual.getServerImage(), is("test/mcp-llm-runtime:test"));
assertThat(actual.getBaseServerImage(), is("test/llama.cpp:test"));
assertThat(actual.getBaseServerImageDigest(),
is("test-base-server-image-digest"));
@@ -241,7 +233,7 @@ class LLME2EConfigurationTest {
private LLME2EConfiguration createConfiguration(final RuntimeMode
runtimeMode, final Path artifactRoot) {
return new LLME2EConfiguration("http://127.0.0.1:8080/v1";,
"openai-compatible", "ggml-org/Qwen3-1.7B-GGUF:Q4_K_M", "mcp-llm-score", 600,
240, 10,
- artifactRoot, "run-id", runtimeMode, "llama.cpp",
"apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191", "",
+ artifactRoot, "run-id", runtimeMode,
"apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191", "",
new
LLME2EConfiguration.ModelMetadata("ggml-org/Qwen3-1.7B-GGUF",
"Qwen3-1.7B-Q4_K_M.gguf", "Q4_K_M", "daeb8e2d528a760970442092f6bf1e55c3b659eb",
"configured-model-sha256"));
}
diff --git
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/conversation/client/LLMChatModelClientTest.java
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/conversation/client/LLMChatModelClientTest.java
index 857b4219f93..36b844e27be 100644
---
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/conversation/client/LLMChatModelClientTest.java
+++
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/conversation/client/LLMChatModelClientTest.java
@@ -162,7 +162,7 @@ class LLMChatModelClientTest {
private LLME2EConfiguration createConfiguration(final String baseUrl,
final int readyTimeoutSeconds) {
return new LLME2EConfiguration(baseUrl, "openai-compatible",
REQUIRED_MODEL, "mcp-llm-score", readyTimeoutSeconds, 30, 10,
- Path.of("target/llm-e2e"), "run-id", RuntimeMode.DOCKER,
"llama.cpp", "apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191",
+ Path.of("target/llm-e2e"), "run-id", RuntimeMode.DOCKER,
"apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191",
"test-base-server-image-digest", MODEL_METADATA);
}
diff --git
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupport.java
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupport.java
index cdf025374a4..94288ab8205 100644
---
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupport.java
+++
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupport.java
@@ -195,7 +195,7 @@ public final class LLMRuntimeSupport {
result.put("runtimeMode", config.getRuntimeMode().getValue());
result.put("dockerOwned", true);
result.put("provider", config.getModelProvider());
- result.put("serverRuntime", config.getServerRuntime());
+ result.put("serverRuntime", "llama.cpp");
result.put("serverImage", config.getServerImage());
result.put("serverImageId", serverImageId);
result.put("baseServerImage", config.getBaseServerImage());
@@ -225,7 +225,6 @@ public final class LLMRuntimeSupport {
&& configuration.getRuntimeMode() ==
config.getRuntimeMode()
&&
configuration.getModelName().equals(config.getModelName())
&& configuration.getApiKey().equals(config.getApiKey())
- &&
configuration.getServerRuntime().equals(config.getServerRuntime())
&&
configuration.getServerImage().equals(config.getServerImage())
&&
configuration.getBaseServerImage().equals(config.getBaseServerImage())
&&
configuration.getBaseServerImageDigest().equals(config.getBaseServerImageDigest())
diff --git
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupportTest.java
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupportTest.java
index 4dc2eb67e22..c6219da67dd 100644
---
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupportTest.java
+++
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/fixture/LLMRuntimeSupportTest.java
@@ -85,7 +85,7 @@ class LLMRuntimeSupportTest {
@Test
void assertPrepareWithUnsupportedProvider() {
LLME2EConfiguration config = new
LLME2EConfiguration("http://127.0.0.1:8080/v1";, "openai", REQUIRED_MODEL,
"mcp-llm-score", 600, 240, 10,
- Path.of("target/llm-e2e"), "run-id", RuntimeMode.DOCKER,
"llama.cpp", "apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191",
+ Path.of("target/llm-e2e"), "run-id", RuntimeMode.DOCKER,
"apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191",
"test-base-server-image-digest", MODEL_METADATA);
IllegalStateException actualException =
assertThrows(IllegalStateException.class, () ->
LLMRuntimeSupport.prepare(config));
assertThat(actualException.getMessage(), is("MCP LLM E2E requires
provider openai-compatible."));
@@ -93,7 +93,7 @@ class LLMRuntimeSupportTest {
private LLME2EConfiguration createConfiguration(final RuntimeMode
runtimeMode, final String modelName, final String baseUrl) {
return new LLME2EConfiguration(baseUrl, "openai-compatible",
modelName, "mcp-llm-score", 2, 2, 10, Path.of("target/llm-e2e"), "run-id",
- runtimeMode, "llama.cpp",
"apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191", "test-base-server-image-digest",
MODEL_METADATA);
+ runtimeMode, "apache/shardingsphere-mcp-llm-runtime:local",
"ghcr.io/ggml-org/llama.cpp:server-b9191", "test-base-server-image-digest",
MODEL_METADATA);
}
private HttpServer startModelServer(final String modelName) throws
IOException {
diff --git
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/suite/usability/LLMUsabilitySuiteRunnerTest.java
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/suite/usability/LLMUsabilitySuiteRunnerTest.java
index c9b77a18ce0..fc379e94338 100644
---
a/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/suite/usability/LLMUsabilitySuiteRunnerTest.java
+++
b/test/e2e/mcp/src/test/java/org/apache/shardingsphere/test/e2e/mcp/llm/suite/usability/LLMUsabilitySuiteRunnerTest.java
@@ -86,6 +86,6 @@ class LLMUsabilitySuiteRunnerTest {
private LLME2EConfiguration createConfiguration() {
return new LLME2EConfiguration("http://127.0.0.1:8080/v1";, "provider",
"model", "api-key", 1, 1, 1, tempDir, "run-id", RuntimeMode.EXTERNAL_DEBUG,
- "runtime", "server-image", "base-image", "", new
LLME2EConfiguration.ModelMetadata("repository", "model.gguf", "Q4", "revision",
"sha256"));
+ "server-image", "base-image", "", new
LLME2EConfiguration.ModelMetadata("repository", "model.gguf", "Q4", "revision",
"sha256"));
}
}
diff --git a/test/e2e/mcp/src/test/resources/docker/llm-runtime/Dockerfile
b/test/e2e/mcp/src/test/resources/docker/llm-runtime/Dockerfile
index ac58a860b61..973810850c8 100644
--- a/test/e2e/mcp/src/test/resources/docker/llm-runtime/Dockerfile
+++ b/test/e2e/mcp/src/test/resources/docker/llm-runtime/Dockerfile
@@ -17,7 +17,6 @@
ARG BASE_IMAGE
FROM ${BASE_IMAGE}
-ARG SERVER_RUNTIME
ARG MODEL_REPOSITORY
ARG MODEL_QUANTIZATION
ARG MODEL_REFERENCE
@@ -28,7 +27,7 @@ ARG MODEL_SHA256
LABEL org.opencontainers.image.title="Apache ShardingSphere MCP LLM E2E
Runtime"
LABEL org.opencontainers.image.description="Prepackaged model server for MCP
LLM E2E score evidence"
LABEL org.opencontainers.image.licenses="Apache-2.0"
-LABEL org.apache.shardingsphere.mcp.llm.server-runtime="${SERVER_RUNTIME}"
+LABEL org.apache.shardingsphere.mcp.llm.runtime="llama.cpp"
LABEL org.apache.shardingsphere.mcp.llm.model-repository="${MODEL_REPOSITORY}"
LABEL
org.apache.shardingsphere.mcp.llm.model-quantization="${MODEL_QUANTIZATION}"
LABEL org.apache.shardingsphere.mcp.llm.model-reference="${MODEL_REFERENCE}"
diff --git a/test/e2e/mcp/src/test/resources/docker/llm-runtime/build-local.sh
b/test/e2e/mcp/src/test/resources/docker/llm-runtime/build-local.sh
index 20d0bf42a86..ee18a69cddd 100644
--- a/test/e2e/mcp/src/test/resources/docker/llm-runtime/build-local.sh
+++ b/test/e2e/mcp/src/test/resources/docker/llm-runtime/build-local.sh
@@ -100,7 +100,6 @@ if [ -z "${TARGET_PLATFORM}" ]; then
fi
BASE_IMAGE="${MCP_LLM_BASE_SERVER_IMAGE:-$(read_required_property
"mcp.llm.base-server-image")}"
BASE_DIGEST="${MCP_LLM_BASE_SERVER_IMAGE_DIGEST:-$(read_optional_property
"mcp.llm.base-server-image-digest")}"
-SERVER_RUNTIME="${MCP_LLM_SERVER_RUNTIME:-$(read_required_property
"mcp.llm.server-runtime")}"
MODEL_REPOSITORY="${MCP_LLM_MODEL_REPOSITORY:-$(read_required_property
"mcp.llm.model-repository")}"
MODEL_QUANTIZATION="${MCP_LLM_MODEL_QUANTIZATION:-$(read_required_property
"mcp.llm.model-quantization")}"
MODEL_REFERENCE="${MCP_LLM_MODEL:-$(read_required_property "mcp.llm.model")}"
@@ -121,7 +120,6 @@ esac
if [ "--dry-run" = "${MODE}" ] || [ "--print" = "${MODE}" ]; then
echo "target_platform=${TARGET_PLATFORM}"
echo "base_image=${BASE_IMAGE}"
- echo "server_runtime=${SERVER_RUNTIME}"
echo "model_repository=${MODEL_REPOSITORY}"
echo "model_quantization=${MODEL_QUANTIZATION}"
echo "model_reference=${MODEL_REFERENCE}"
@@ -142,7 +140,6 @@ fi
docker build \
--platform "${TARGET_PLATFORM}" \
--build-arg "BASE_IMAGE=${BASE_IMAGE}" \
- --build-arg "SERVER_RUNTIME=${SERVER_RUNTIME}" \
--build-arg "MODEL_REPOSITORY=${MODEL_REPOSITORY}" \
--build-arg "MODEL_QUANTIZATION=${MODEL_QUANTIZATION}" \
--build-arg "MODEL_REFERENCE=${MODEL_REFERENCE}" \
diff --git a/test/e2e/mcp/src/test/resources/env/e2e-env.properties
b/test/e2e/mcp/src/test/resources/env/e2e-env.properties
index 3c3424d527b..104cd8b163b 100644
--- a/test/e2e/mcp/src/test/resources/env/e2e-env.properties
+++ b/test/e2e/mcp/src/test/resources/env/e2e-env.properties
@@ -35,7 +35,6 @@ mcp.llm.request-timeout-seconds=240
mcp.llm.max-turns=10
mcp.llm.artifact-root=target/llm-e2e
mcp.llm.run-id=
-mcp.llm.server-runtime=llama.cpp
mcp.llm.server-image=apache/shardingsphere-mcp-llm-runtime:local
mcp.llm.target-platform=
mcp.llm.base-server-image=ghcr.io/ggml-org/llama.cpp:server-b9191
