[CI] Optimize CI time (#4821)

### What this PR does / why we need it?
Considering that long queues severely impact the developer experience,
we have decided to make the following changes:
1. Changes will use the self_hosted runner
2. e2e-2card will use the A3 node.

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
Co-authored-by: Mengqing Cao <cmq0113@163.com>

.github/workflows/_e2e_test.yaml

@@ -112,9 +112,9 @@ jobs:
 
   e2e-2-cards:
     name: multicard-2
-    runs-on: ${{ inputs.runner }}-2
+    runs-on: linux-aarch64-a3-2
     container:
-      image: ${{ inputs.image }}
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-a3-ubuntu22.04-py3.11
       env:
         VLLM_LOGGING_LEVEL: ERROR
         VLLM_USE_MODELSCOPE: True
@@ -126,9 +126,11 @@ jobs:
 
       - name: Config mirrors
         run: |
-          sed -Ei 's@(ports|archive).ubuntu.com@cache-service.nginx-pypi-cache.svc.cluster.local:8081@g' /etc/apt/sources.list
+          # Fix me: use nginx cache rather than the pypi
-          pip config set global.index-url http://cache-service.nginx-pypi-cache.svc.cluster.local/pypi/simple
+          # sed -Ei 's@(ports|archive).ubuntu.com@cache-service.nginx-pypi-cache.svc.cluster.local:8081@g' /etc/apt/sources.list
-          pip config set global.trusted-host cache-service.nginx-pypi-cache.svc.cluster.local
+          # pip config set global.index-url http://cache-service.nginx-pypi-cache.svc.cluster.local/pypi/simple
+          # pip config set global.trusted-host cache-service.nginx-pypi-cache.svc.cluster.local
+          pip config set global.index-url https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple
           apt-get update -y
           apt install git -y
 

.github/workflows/pre-commit.yml

@@ -40,4 +40,3 @@ jobs:
         SHELLCHECK_OPTS: "--exclude=SC2046,SC2006,SC2086" # Exclude SC2046, SC2006, SC2086 for actionlint
       with:
         extra_args: --all-files --hook-stage manual
-

.github/workflows/vllm_ascend_test_pr_full.yaml

@@ -38,13 +38,14 @@ concurrency:
 
 jobs:
   changes:
-    runs-on: ubuntu-latest
+    runs-on: linux-aarch64-a2-0
     if: ${{ contains(github.event.pull_request.labels.*.name, 'ready') && contains(github.event.pull_request.labels.*.name, 'ready-for-test') }}
     outputs:
       e2e_tracker: ${{ steps.filter.outputs.e2e_tracker }}
       ut_tracker: ${{ steps.filter.outputs.ut_tracker }}
     steps:
-      - uses: actions/checkout@v6.0.1
+      # NOTE: Do not update the version of checkout, there have some issue on self_hosted runner with the higher version
+      - uses: actions/checkout@v4
       - uses: dorny/paths-filter@v3
         id: filter
         with:

.github/workflows/vllm_ascend_test_pr_light.yaml

@@ -44,12 +44,13 @@ jobs:
     with:
       vllm: ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
   changes:
-    runs-on: ubuntu-latest
+    runs-on: linux-aarch64-a2-0
     outputs:
       e2e_tracker: ${{ steps.filter.outputs.e2e_tracker }}
       ut_tracker: ${{ steps.filter.outputs.ut_tracker }}
     steps:
-      - uses: actions/checkout@v6.0.1
+      # NOTE: Do not update the version of checkout, there have some issue on self_hosted runner with the higher version
+      - uses: actions/checkout@v4
       - uses: dorny/paths-filter@v3
         id: filter
         with: