Slim down docker cache size (#3190)

* slim down docker cache size

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* remove old centos images

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* troubleshoot test failure

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* fix wget version ref

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* refactor caching mechanisms

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* add cache cleanup steps

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* simplify deleting cache

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* fix first clone issue

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* add tool dep

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

---------

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
This commit is contained in:
Alex Goodman
2024-09-09 11:15:13 -04:00
committed by GitHub
parent deabd4115a
commit 0a3f513f92
75 changed files with 1302 additions and 382 deletions
+14 -9
View File
@@ -13,16 +13,15 @@ inputs:
cache-key-prefix:
description: "Prefix all cache keys with this value"
required: true
default: "1ac8281053"
compute-fingerprints:
description: "Compute test fixture fingerprints"
default: "181053ac82"
download-test-fixture-cache:
description: "Download test fixture cache from OCI and github actions"
required: true
default: "true"
default: "false"
bootstrap-apt-packages:
description: "Space delimited list of tools to install via apt"
default: "libxml2-utils"
runs:
using: "composite"
steps:
@@ -54,8 +53,14 @@ runs:
run: |
DEBIAN_FRONTEND=noninteractive sudo apt update && sudo -E apt install -y ${{ inputs.bootstrap-apt-packages }}
- name: Create all cache fingerprints
if: inputs.compute-fingerprints == 'true'
shell: bash
run: make fingerprints
- name: Restore ORAS cache from github actions
if: inputs.download-test-fixture-cache == 'true'
uses: actions/cache@704facf57e6136b1bc63b828d79edcd491f0ee84 # v3.3.2
with:
path: ${{ github.workspace }}/.tmp/oras-cache
key: ${{ inputs.cache-key-prefix }}-oras-cache
- name: Download test fixture cache
if: inputs.download-test-fixture-cache == 'true'
shell: bash
run: make download-test-fixture-cache
-11
View File
@@ -1,11 +0,0 @@
#!/usr/bin/env bash
red=$(tput setaf 1)
bold=$(tput bold)
normal=$(tput sgr0)
# assert we are running in CI (or die!)
if [[ -z "$CI" ]]; then
echo "${bold}${red}This step should ONLY be run in CI. Exiting...${normal}"
exit 1
fi
+135
View File
@@ -0,0 +1,135 @@
#!/usr/bin/env python3
from __future__ import annotations
import os
import glob
import sys
import json
import hashlib
IGNORED_PREFIXES = []
def find_fingerprints_and_check_dirs(base_dir):
all_fingerprints = set(glob.glob(os.path.join(base_dir, '**', 'test*', '**', '*.fingerprint'), recursive=True))
all_fingerprints = {os.path.relpath(fp) for fp in all_fingerprints
if not any(fp.startswith(prefix) for prefix in IGNORED_PREFIXES)}
if not all_fingerprints:
show("No .fingerprint files or cache directories found.")
exit(1)
missing_content = []
valid_paths = set()
fingerprint_contents = []
for fingerprint in all_fingerprints:
path = fingerprint.replace('.fingerprint', '')
if not os.path.exists(path):
missing_content.append(path)
continue
if not os.path.isdir(path):
valid_paths.add(path)
continue
if os.listdir(path):
valid_paths.add(path)
else:
missing_content.append(path)
with open(fingerprint, 'r') as f:
content = f.read().strip()
fingerprint_contents.append((fingerprint, content))
return sorted(valid_paths), missing_content, fingerprint_contents
def parse_fingerprint_contents(fingerprint_content):
input_map = {}
for line in fingerprint_content.splitlines():
digest, path = line.split()
input_map[path] = digest
return input_map
def calculate_sha256(fingerprint_contents):
sorted_fingerprint_contents = sorted(fingerprint_contents, key=lambda x: x[0])
concatenated_contents = ''.join(content for _, content in sorted_fingerprint_contents)
sha256_hash = hashlib.sha256(concatenated_contents.encode()).hexdigest()
return sha256_hash
def calculate_file_sha256(file_path):
sha256_hash = hashlib.sha256()
with open(file_path, 'rb') as f:
for byte_block in iter(lambda: f.read(4096), b""):
sha256_hash.update(byte_block)
return sha256_hash.hexdigest()
def show(*s: str):
print(*s, file=sys.stderr)
def main(file_path: str | None):
base_dir = '.'
valid_paths, missing_content, fingerprint_contents = find_fingerprints_and_check_dirs(base_dir)
if missing_content:
show("The following paths are missing or have no content, but have corresponding .fingerprint files:")
for path in sorted(missing_content):
show(f"- {path}")
show("Please ensure these paths exist and have content if they are directories.")
exit(1)
sha256_hash = calculate_sha256(fingerprint_contents)
paths_with_digests = []
for path in sorted(valid_paths):
fingerprint_file = f"{path}.fingerprint"
try:
if os.path.exists(fingerprint_file):
file_digest = calculate_file_sha256(fingerprint_file)
# Parse the fingerprint file to get the digest/path tuples
with open(fingerprint_file, 'r') as f:
fingerprint_content = f.read().strip()
input_map = parse_fingerprint_contents(fingerprint_content)
paths_with_digests.append({
"path": path,
"digest": file_digest,
"input": input_map
})
except Exception as e:
show(f"Error processing {fingerprint_file}: {e}")
raise e
output = {
"digest": sha256_hash,
"paths": paths_with_digests
}
content = json.dumps(output, indent=2, sort_keys=True)
if file_path:
with open(file_path, 'w') as f:
f.write(content)
print(content)
if __name__ == "__main__":
file_path = None
if len(sys.argv) > 1:
file_path = sys.argv[1]
main(file_path)
+70
View File
@@ -0,0 +1,70 @@
#!/usr/bin/env python3
import os
import subprocess
import hashlib
BOLD = '\033[1m'
YELLOW = '\033[0;33m'
RESET = '\033[0m'
def print_message(message):
print(f"{YELLOW}{message}{RESET}")
def sha256sum(filepath):
h = hashlib.sha256()
with open(filepath, 'rb') as f:
for chunk in iter(lambda: f.read(4096), b""):
h.update(chunk)
return h.hexdigest()
def is_git_tracked_or_untracked(directory):
"""Returns a sorted list of files in the directory that are tracked or not ignored by Git."""
result = subprocess.run(
["git", "ls-files", "--cached", "--others", "--exclude-standard"],
cwd=directory,
stdout=subprocess.PIPE,
text=True
)
return sorted(result.stdout.strip().splitlines())
def find_test_fixture_dirs_with_images(base_dir):
"""Find directories that contain 'test-fixtures' and at least one 'image-*' directory."""
for root, dirs, files in os.walk(base_dir):
if 'test-fixtures' in root:
image_dirs = [d for d in dirs if d.startswith('image-')]
if image_dirs:
yield os.path.realpath(root)
def generate_fingerprints():
print_message("creating fingerprint files for docker fixtures...")
for test_fixture_dir in find_test_fixture_dirs_with_images('.'):
cache_fingerprint_path = os.path.join(test_fixture_dir, 'cache.fingerprint')
with open(cache_fingerprint_path, 'w') as fingerprint_file:
for image_dir in find_image_dirs(test_fixture_dir):
for file in is_git_tracked_or_untracked(image_dir):
file_path = os.path.join(image_dir, file)
checksum = sha256sum(file_path)
path_from_fixture_dir = os.path.relpath(file_path, test_fixture_dir)
fingerprint_file.write(f"{checksum} {path_from_fixture_dir}\n")
def find_image_dirs(test_fixture_dir):
"""Find all 'image-*' directories inside a given test-fixture directory."""
result = []
for root, dirs, files in os.walk(test_fixture_dir):
for dir_name in dirs:
if dir_name.startswith('image-'):
result.append(os.path.join(root, dir_name))
return sorted(result)
if __name__ == "__main__":
generate_fingerprints()
Regular → Executable
+2
View File
@@ -1,3 +1,5 @@
#!/usr/bin/env python3
from __future__ import annotations
import sys
Regular → Executable
+2
View File
@@ -1,3 +1,5 @@
#!/usr/bin/env python3
import unittest
from unittest.mock import patch
import subprocess
+1 -1
View File
@@ -1,4 +1,4 @@
name: "Release"
name: "Release: version file"
on:
@@ -0,0 +1,39 @@
name: "Test fixture cache: publish"
on:
workflow_dispatch:
schedule:
# run nightly at 4AM UTC
- cron: "0 4 * * *"
permissions:
contents: read
jobs:
Publish:
name: "Publish test fixture image cache"
# we use this runner to get enough storage space for docker images and fixture cache
runs-on: ubuntu-22.04-4core-16gb
permissions:
packages: write
steps:
- uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 #v4.1.7
- name: Bootstrap environment
uses: ./.github/actions/bootstrap
with:
# we want to rebuild the cache with no previous state
download-test-fixture-cache: false
- name: Run all tests
run: make test
env:
# we want to rebuild the cache with no previous state
DOWNLOAD_TEST_FIXTURE_CACHE: "false"
- name: Login to GitHub Container Registry (ORAS)
run: echo "${{ secrets.GITHUB_TOKEN }}" | .tool/oras login ghcr.io -u ${{ github.actor }} --password-stdin
- name: Publish test fixture cache
run: make upload-test-fixture-cache
@@ -19,7 +19,6 @@ jobs:
uses: ./.github/actions/bootstrap
with:
bootstrap-apt-packages: ""
compute-fingerprints: "false"
go-dependencies: false
- name: "Update tool versions"
+26 -67
View File
@@ -35,48 +35,8 @@ jobs:
- name: Bootstrap environment
uses: ./.github/actions/bootstrap
- name: Restore file executable test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/file/cataloger/executable/test-fixtures/elf/bin
key: ${{ runner.os }}-unit-file-executable-elf-cache-${{ hashFiles( 'syft/file/cataloger/executable/test-fixtures/elf/cache.fingerprint' ) }}
- name: Restore file executable shared-info test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/file/cataloger/executable/test-fixtures/shared-info/bin
key: ${{ runner.os }}-unit-file-executable-shared-info-cache-${{ hashFiles( 'syft/file/cataloger/executable/test-fixtures/shared-info/cache.fingerprint' ) }}
- name: Restore Java test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/pkg/cataloger/java/test-fixtures/java-builds/packages
key: ${{ runner.os }}-unit-java-cache-${{ hashFiles( 'syft/pkg/cataloger/java/test-fixtures/java-builds/cache.fingerprint' ) }}
- name: Restore RPM test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/pkg/cataloger/redhat/test-fixtures/rpms
key: ${{ runner.os }}-unit-rpm-cache-${{ hashFiles( 'syft/pkg/cataloger/redhat/test-fixtures/rpms.fingerprint' ) }}
- name: Restore go binary test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/pkg/cataloger/golang/test-fixtures/archs/binaries
key: ${{ runner.os }}-unit-go-binaries-cache-${{ hashFiles( 'syft/pkg/cataloger/golang/test-fixtures/archs/binaries.fingerprint' ) }}
- name: Restore binary cataloger test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/pkg/cataloger/binary/test-fixtures/classifiers/bin
key: ${{ runner.os }}-unit-binary-cataloger-cache-${{ hashFiles( 'syft/pkg/cataloger/binary/test-fixtures/cache.fingerprint' ) }}
- name: Restore Kernel test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: syft/pkg/cataloger/kernel/test-fixtures/cache
key: ${{ runner.os }}-unit-kernel-cache-${{ hashFiles( 'syft/pkg/cataloger/kernel/test-fixtures/cache.fingerprint' ) }}
download-test-fixture-cache: true
- name: Run unit tests
run: make unit
@@ -91,16 +51,12 @@ jobs:
- name: Bootstrap environment
uses: ./.github/actions/bootstrap
with:
download-test-fixture-cache: true
- name: Validate syft output against the CycloneDX schema
run: make validate-cyclonedx-schema
- name: Restore integration test cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: ${{ github.workspace }}/cmd/syft/internal/test/integration/test-fixtures/cache
key: ${{ runner.os }}-integration-test-cache-${{ hashFiles('/cmd/syft/internal/test/integration/test-fixtures/cache.fingerprint') }}
- name: Run integration tests
run: make integration
@@ -143,6 +99,8 @@ jobs:
- name: Bootstrap environment
uses: ./.github/actions/bootstrap
with:
download-test-fixture-cache: true
- name: Download snapshot build
id: snapshot-cache
@@ -162,13 +120,6 @@ jobs:
- name: Run comparison tests (Linux)
run: make compare-linux
- name: Restore install.sh test image cache
id: install-test-image-cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: ${{ github.workspace }}/test/install/cache
key: ${{ runner.os }}-install-test-image-cache-${{ hashFiles('test/install/cache.fingerprint') }}
- name: Load test image cache
if: steps.install-test-image-cache.outputs.cache-hit == 'true'
run: make install-test-cache-load
@@ -196,8 +147,8 @@ jobs:
uses: ./.github/actions/bootstrap
with:
bootstrap-apt-packages: ""
compute-fingerprints: "false"
go-dependencies: false
download-test-fixture-cache: true
- name: Download snapshot build
id: snapshot-cache
@@ -214,13 +165,6 @@ jobs:
if: steps.snapshot-cache.outputs.cache-hit != 'true'
run: echo "unable to download snapshots from previous job" && false
- name: Restore docker image cache for compare testing
id: mac-compare-testing-cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: image.tar
key: ${{ runner.os }}-${{ hashFiles('test/compare/mac.sh') }}
- name: Run comparison tests (Mac)
run: make compare-mac
@@ -238,12 +182,8 @@ jobs:
- name: Bootstrap environment
uses: ./.github/actions/bootstrap
- name: Restore CLI test-fixture cache
uses: actions/cache@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 #v4.0.2
with:
path: ${{ github.workspace }}/test/cli/test-fixtures/cache
key: ${{ runner.os }}-cli-test-cache-${{ hashFiles('test/cli/test-fixtures/cache.fingerprint') }}
download-test-fixture-cache: true
- name: Download snapshot build
id: snapshot-cache
@@ -262,3 +202,22 @@ jobs:
- name: Run CLI Tests (Linux)
run: make cli
Cleanup-Cache:
name: "Cleanup snapshot cache"
if: always()
runs-on: ubuntu-20.04
permissions:
actions: write
needs:
- Acceptance-Linux
- Acceptance-Mac
- Cli-Linux
steps:
- uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 #v4.1.7
- name: Delete snapshot cache
run: gh cache delete "snapshot-build-${{ github.run_id }}"
env:
GH_TOKEN: ${{ github.token }}