diff --git a/docker/Makefile b/docker/Makefile
index 8432710af43b..b95ea971ef37 100644
--- a/docker/Makefile
+++ b/docker/Makefile
@@ -1,6 +1,8 @@
 # Default base image for the docker build as defined in Dockerfile.multi
 BASE_IMAGE         ?= $(shell grep '^ARG BASE_IMAGE=' Dockerfile.multi | grep -o '=.*' | tr -d '="')
 BASE_TAG           ?= $(shell grep '^ARG BASE_TAG=' Dockerfile.multi | grep -o '=.*' | tr -d '="')
+TRITON_IMAGE       ?= $(shell grep '^ARG TRITON_IMAGE=' Dockerfile.multi | grep -o '=.*' | tr -d '="')
+TRITON_BASE_TAG    ?= $(shell grep '^ARG TRITON_BASE_TAG=' Dockerfile.multi | grep -o '=.*' | tr -d '="')
 # Name of the new image
 IMAGE_NAME         ?= tensorrt_llm
 IMAGE_TAG          ?= latest
@@ -80,6 +82,8 @@ endef
 		--progress $(DOCKER_PROGRESS) \
 		$(if $(BASE_IMAGE), --build-arg BASE_IMAGE=$(BASE_IMAGE)) \
 		$(if $(BASE_TAG), --build-arg BASE_TAG=$(BASE_TAG)) \
+		$(if $(TRITON_IMAGE), --build-arg TRITON_IMAGE=$(TRITON_IMAGE)) \
+		$(if $(TRITON_BASE_TAG), --build-arg TRITON_BASE_TAG=$(TRITON_BASE_TAG)) \
 		$(if $(BUILD_WHEEL_ARGS), --build-arg BUILD_WHEEL_ARGS="$(BUILD_WHEEL_ARGS)") \
 		$(if $(BUILD_WHEEL_SCRIPT), --build-arg BUILD_WHEEL_SCRIPT="$(BUILD_WHEEL_SCRIPT)") \
 		$(if $(TORCH_INSTALL_TYPE), --build-arg TORCH_INSTALL_TYPE="$(TORCH_INSTALL_TYPE)") \
@@ -187,16 +191,16 @@ jenkins-aarch64_%: STAGE = tritondevel
 jenkins-rockylinux8_%: PYTHON_VERSION_TAG_ID = $(if $(findstring 3.12,${PYTHON_VERSION}),PY312,$(if $(findstring 3.10,${PYTHON_VERSION}),PY310,$(error Unknown PYTHON_VERSION specified)))
 jenkins-rockylinux8_%: IMAGE_WITH_TAG = $(shell . ../jenkins/current_image_tags.properties && echo $$LLM_ROCKYLINUX8_${PYTHON_VERSION_TAG_ID}_DOCKER_IMAGE)
 jenkins-rockylinux8_%: STAGE = tritondevel
-jenkins-rockylinux8_%: BASE_IMAGE = nvidia/cuda
+jenkins-rockylinux8_%: BASE_IMAGE = nvcr.io/nvidia/cuda
 jenkins-rockylinux8_%: BASE_TAG = 12.9.1-devel-rockylinux8
 
 rockylinux8_%: STAGE = tritondevel
-rockylinux8_%: BASE_IMAGE = nvidia/cuda
+rockylinux8_%: BASE_IMAGE = nvcr.io/nvidia/cuda
 rockylinux8_%: BASE_TAG = 12.9.1-devel-rockylinux8
 
 # For x86_64 and aarch64
 ubuntu22_%: STAGE = tritondevel
-ubuntu22_%: BASE_IMAGE = nvidia/cuda
+ubuntu22_%: BASE_IMAGE = nvcr.io/nvidia/cuda
 ubuntu22_%: BASE_TAG = 12.9.1-devel-ubuntu22.04
 
 trtllm_%: STAGE = release
diff --git a/jenkins/BuildDockerImage.groovy b/jenkins/BuildDockerImage.groovy
index 5aa61708f5bf..fea7d8261c97 100644
--- a/jenkins/BuildDockerImage.groovy
+++ b/jenkins/BuildDockerImage.groovy
@@ -281,23 +281,30 @@ def buildImage(config, imageKeyToTag)
     try {
         def build_jobs = BUILD_JOBS
         // Fix the triton image pull timeout issue
+        def BASE_IMAGE = sh(script: "cd ${LLM_ROOT} && grep 'ARG BASE_IMAGE=' docker/Dockerfile.multi | grep -o '=.*' | tr -d '=\"'", returnStdout: true).trim()
         def TRITON_IMAGE = sh(script: "cd ${LLM_ROOT} && grep 'ARG TRITON_IMAGE=' docker/Dockerfile.multi | grep -o '=.*' | tr -d '=\"'", returnStdout: true).trim()
         def TRITON_BASE_TAG = sh(script: "cd ${LLM_ROOT} && grep 'ARG TRITON_BASE_TAG=' docker/Dockerfile.multi | grep -o '=.*' | tr -d '=\"'", returnStdout: true).trim()
 
+        if (target == "rockylinux8") {
+            BASE_IMAGE = sh(script: "cd ${LLM_ROOT} && grep 'jenkins-rockylinux8_%: BASE_IMAGE =' docker/Makefile | grep -o '=.*' | tr -d '=\"'", returnStdout: true).trim()
+        }
+
+        // Replace the base image and triton image with the internal mirror
+        BASE_IMAGE = BASE_IMAGE.replace("nvcr.io/", "urm.nvidia.com/docker/")
+        TRITON_IMAGE = TRITON_IMAGE.replace("nvcr.io/", "urm.nvidia.com/docker/")
+
         if (dependent) {
             stage ("make ${dependent.target}_${action} (${arch})") {
-                retry(3) {
-                    sh "docker pull ${TRITON_IMAGE}:${TRITON_BASE_TAG}"
-                }
-                retry(3) {
-                    sh """
-                    cd ${LLM_ROOT} && make -C docker ${dependent.target}_${action} \
-                    TORCH_INSTALL_TYPE=${torchInstallType} \
-                    IMAGE_WITH_TAG=${dependentImageWithTag} \
-                    STAGE=${dependent.dockerfileStage} \
-                    BUILD_WHEEL_OPTS='-j ${build_jobs}' ${args}
-                    """
-                }
+                trtllm_utils.llmExecStepWithRetry(this, script: "docker pull ${TRITON_IMAGE}:${TRITON_BASE_TAG}", sleepInSecs: 300, shortCommondRunTimeMax: 7200)
+                trtllm_utils.llmExecStepWithRetry(this, script: """
+                cd ${LLM_ROOT} && make -C docker ${dependent.target}_${action} \
+                BASE_IMAGE=${BASE_IMAGE} \
+                TRITON_IMAGE=${TRITON_IMAGE} \
+                TORCH_INSTALL_TYPE=${torchInstallType} \
+                IMAGE_WITH_TAG=${dependentImageWithTag} \
+                STAGE=${dependent.dockerfileStage} \
+                BUILD_WHEEL_OPTS='-j ${build_jobs}' ${args}
+                """, sleepInSecs: 300, shortCommondRunTimeMax: 7200)
                 args += " DEVEL_IMAGE=${dependentImageWithTag}"
                 if (target == "ngc-release") {
                     imageKeyToTag["NGC Devel Image ${config.arch}"] = dependentImageWithTag
@@ -315,18 +322,16 @@ def buildImage(config, imageKeyToTag)
             }
         }
         stage ("make ${target}_${action} (${arch})") {
-            retry(3) {
-                sh "docker pull ${TRITON_IMAGE}:${TRITON_BASE_TAG}"
-            }
-            retry(3) {
-                sh """
-                cd ${LLM_ROOT} && make -C docker ${target}_${action} \
-                TORCH_INSTALL_TYPE=${torchInstallType} \
-                IMAGE_WITH_TAG=${imageWithTag} \
-                STAGE=${dockerfileStage} \
-                BUILD_WHEEL_OPTS='-j ${build_jobs}' ${args}
-                """
-            }
+            trtllm_utils.llmExecStepWithRetry(this, script: "docker pull ${TRITON_IMAGE}:${TRITON_BASE_TAG}", sleepInSecs: 300, shortCommondRunTimeMax: 7200)
+            trtllm_utils.llmExecStepWithRetry(this, script: """
+            cd ${LLM_ROOT} && make -C docker ${target}_${action} \
+            BASE_IMAGE=${BASE_IMAGE} \
+            TRITON_IMAGE=${TRITON_IMAGE} \
+            TORCH_INSTALL_TYPE=${torchInstallType} \
+            IMAGE_WITH_TAG=${imageWithTag} \
+            STAGE=${dockerfileStage} \
+            BUILD_WHEEL_OPTS='-j ${build_jobs}' ${args}
+            """, sleepInSecs: 300, shortCommondRunTimeMax: 7200)
             if (target == "ngc-release") {
                 imageKeyToTag["NGC Release Image ${config.arch}"] = imageWithTag
             }
@@ -336,6 +341,8 @@ def buildImage(config, imageKeyToTag)
             stage ("custom tag: ${customTag} (${arch})") {
                 sh """
                 cd ${LLM_ROOT} && make -C docker ${target}_${action} \
+                BASE_IMAGE=${BASE_IMAGE} \
+                TRITON_IMAGE=${TRITON_IMAGE} \
                 TORCH_INSTALL_TYPE=${torchInstallType} \
                 IMAGE_WITH_TAG=${customImageWithTag} \
                 STAGE=${dockerfileStage} \
diff --git a/jenkins/L0_Test.groovy b/jenkins/L0_Test.groovy
index b40c7a11a7e4..22a556936066 100644
--- a/jenkins/L0_Test.groovy
+++ b/jenkins/L0_Test.groovy
@@ -39,7 +39,7 @@ LLM_ROCKYLINUX8_PY310_DOCKER_IMAGE = env.wheelDockerImagePy310
 LLM_ROCKYLINUX8_PY312_DOCKER_IMAGE = env.wheelDockerImagePy312
 
 // DLFW torch image
-DLFW_IMAGE = "nvcr.io/nvidia/pytorch:25.06-py3"
+DLFW_IMAGE = "urm.nvidia.com/docker/nvidia/pytorch:25.06-py3"
 
 //Ubuntu base image
 UBUNTU_22_04_IMAGE = "urm.nvidia.com/docker/ubuntu:22.04"
diff --git a/jenkins/current_image_tags.properties b/jenkins/current_image_tags.properties
index dee2ee7218f3..ddf5e2aca9d5 100644
--- a/jenkins/current_image_tags.properties
+++ b/jenkins/current_image_tags.properties
@@ -11,7 +11,7 @@
 #
 # NB: Typically, the suffix indicates the PR whose CI pipeline generated the images. In case that
 #     images are adopted from PostMerge pipelines, the abbreviated commit hash is used instead.
-LLM_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-25.06-py3-x86_64-ubuntu24.04-trt10.11.0.33-skip-tritondevel-202508051130-6090
-LLM_SBSA_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-25.06-py3-aarch64-ubuntu24.04-trt10.11.0.33-skip-tritondevel-202508051130-6090
-LLM_ROCKYLINUX8_PY310_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-12.9.1-devel-rocky8-x86_64-rocky8-py310-trt10.11.0.33-skip-tritondevel-202508051130-6090
-LLM_ROCKYLINUX8_PY312_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-12.9.1-devel-rocky8-x86_64-rocky8-py312-trt10.11.0.33-skip-tritondevel-202508051130-6090
+LLM_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-25.06-py3-x86_64-ubuntu24.04-trt10.11.0.33-skip-tritondevel-202508110900-6715
+LLM_SBSA_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:pytorch-25.06-py3-aarch64-ubuntu24.04-trt10.11.0.33-skip-tritondevel-202508110900-6715
+LLM_ROCKYLINUX8_PY310_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-12.9.1-devel-rocky8-x86_64-rocky8-py310-trt10.11.0.33-skip-tritondevel-202508110900-6715
+LLM_ROCKYLINUX8_PY312_DOCKER_IMAGE=urm.nvidia.com/sw-tensorrt-docker/tensorrt-llm:cuda-12.9.1-devel-rocky8-x86_64-rocky8-py312-trt10.11.0.33-skip-tritondevel-202508110900-6715
diff --git a/tests/integration/test_lists/test-db/l0_dgx_h100.yml b/tests/integration/test_lists/test-db/l0_dgx_h100.yml
index 81369dee61b1..6c39e94f63d2 100644
--- a/tests/integration/test_lists/test-db/l0_dgx_h100.yml
+++ b/tests/integration/test_lists/test-db/l0_dgx_h100.yml
@@ -189,5 +189,6 @@ l0_dgx_h100:
     terms:
       stage: post_merge
       backend: triton
+      auto_trigger: others
   tests:
   - triton_server/test_triton_llm.py::test_llmapi_backend[4-0-disableDecoupleMode-tensorrt_llm]
diff --git a/tests/unittest/test_pip_install.py b/tests/unittest/test_pip_install.py
index d75bfbaf4206..11288e09cec0 100644
--- a/tests/unittest/test_pip_install.py
+++ b/tests/unittest/test_pip_install.py
@@ -51,9 +51,6 @@ def test_pip_install():
                         help="The wheel path")
     args = parser.parse_args()
 
-    if not os.environ.get("CUDA_HOME"):
-        os.environ["CUDA_HOME"] = "/usr/local/cuda"
-
     print("##########  Install required system libs  ##########")
     if not os.path.exists("/usr/local/mpi/bin/mpicc"):
         subprocess.check_call("apt-get -y install libopenmpi-dev", shell=True)