[hadoop] 02/06: YARN-7345. GPU Isolation: Incorrect minor device numbers written to devices.deny file. (Jonathan Hung via wangda)

[jh...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

jhung pushed a commit to branch YARN-8200
in repository https://gitbox.apache.org/repos/asf/hadoop.git

commit a9c58d38c1926300029c44f9f13d9cf1093dd650
Author: Wangda Tan <wa...@apache.org>
AuthorDate: Thu Oct 19 14:45:44 2017 -0700

    YARN-7345. GPU Isolation: Incorrect minor device numbers written to devices.deny file. (Jonathan Hung via wangda)
---
 .../native/container-executor/impl/modules/gpu/gpu-module.c |  2 +-
 .../container-executor/test/modules/gpu/test-gpu-module.cc  | 13 +++++++++++++
 2 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/impl/modules/gpu/gpu-module.c b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/impl/modules/gpu/gpu-module.c
index f96645d..1a1b164 100644
--- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/impl/modules/gpu/gpu-module.c
+++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/impl/modules/gpu/gpu-module.c
@@ -108,7 +108,7 @@ static int internal_handle_gpu_request(
     char param_value[128];
     memset(param_value, 0, sizeof(param_value));
     snprintf(param_value, sizeof(param_value), "c %d:%d rwm",
-             major_device_number, i);
+             major_device_number, minor_devices[i]);
 
     int rc = update_cgroups_parameters_func_p("devices", "deny",
       container_id, param_value);
diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/test/modules/gpu/test-gpu-module.cc b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/test/modules/gpu/test-gpu-module.cc
index 7e41fb4..b3d93dc 100644
--- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/test/modules/gpu/test-gpu-module.cc
+++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/native/container-executor/test/modules/gpu/test-gpu-module.cc
@@ -165,6 +165,19 @@ TEST_F(TestGpuModule, test_verify_gpu_module_calls_cgroup_parameter) {
 
   // Verify cgroups parameters
   verify_param_updated_to_cgroups(0, NULL);
+
+  /* Test case 3: block 2 non-sequential devices */
+  cgroups_parameters_invoked.clear();
+  char* argv_2[] = { (char*) "--module-gpu", (char*) "--excluded_gpus", (char*) "1,3",
+                   (char*) "--container_id", container_id };
+  rc = handle_gpu_request(&mock_update_cgroups_parameters,
+     "gpu", 5, argv_2);
+  ASSERT_EQ(0, rc) << "Should success.\n";
+
+  // Verify cgroups parameters
+  const char* expected_cgroups_argv_2[] = { "devices", "deny", container_id, "c 195:1 rwm",
+    "devices", "deny", container_id, "c 195:3 rwm"};
+  verify_param_updated_to_cgroups(8, expected_cgroups_argv_2);
 }
 
 TEST_F(TestGpuModule, test_illegal_cli_parameters) {


---------------------------------------------------------------------
To unsubscribe, e-mail: common-commits-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-commits-help@hadoop.apache.org