Updates UCX to 1.13.1 in Dockerfile-blossom.ubuntu and sets UCX_TLS=^…

…posix (#6573) Signed-off-by: Alessandro Bellina <[email protected]> Signed-off-by: Alessandro Bellina <[email protected]>
NVIDIA · Sep 20, 2022 · 998c76e · 998c76e
1 parent e1cd125
commit 998c76e
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 5 deletions.
diff --git a/jenkins/Dockerfile-blossom.ubuntu b/jenkins/Dockerfile-blossom.ubuntu
@@ -22,12 +22,12 @@
 #       CUDA_VER=11.0+
 #       UBUNTU_VER=18.04 or 20.04
 #       UCX_CUDA_VER=11 (major CUDA version)
-#       UCX_VER=1.12.1
+#       UCX_VER=1.13.1
 ###
 
 ARG CUDA_VER=11.0
 ARG UBUNTU_VER=18.04
-ARG UCX_VER=1.12.1
+ARG UCX_VER=1.13.1
 ARG UCX_CUDA_VER=11
 FROM nvidia/cuda:${CUDA_VER}-runtime-ubuntu${UBUNTU_VER}
 ARG CUDA_VER

diff --git a/jenkins/spark-premerge-build.sh b/jenkins/spark-premerge-build.sh
@@ -109,9 +109,12 @@ rapids_shuffle_smoke_test() {
     }
 
     # using UCX shuffle
-    # Disabled temporarily due to: https://github.com/NVIDIA/spark-rapids/issues/6572
-    # PYSP_TEST_spark_executorEnv_UCX_ERROR_SIGNALS="" \
-    #     invoke_shuffle_integration_test
+    # The UCX_TLS=^posix config is removing posix from the list of memory transports
+    # so that IPC regions are obtained using SysV API instead. This was done because of
+    # itermittent test failures. See: https://github.com/NVIDIA/spark-rapids/issues/6572
+    PYSP_TEST_spark_executorEnv_UCX_ERROR_SIGNALS="" \
+    PYSP_TEST_spark_executorEnv_UCX_TLS="^posix" \
+        invoke_shuffle_integration_test
 
     # using MULTITHREADED shuffle
     PYSP_TEST_spark_rapids_shuffle_mode=MULTITHREADED \