@@ -12,8 +12,9 @@ permissions:
1212
1313jobs :
1414test-go-pg :
15- runs-on :${{ matrix.os == 'macos-latest' && github.repository_owner == 'coder' && 'depot-macos-latest' || matrix.os == 'windows-2022' && github.repository_owner == 'coder' && 'windows-latest-16-cores' || matrix.os }}
16- if :github.ref == 'refs/heads/main'
15+ # make sure to adjust NUM_PARALLEL_PACKAGES and NUM_PARALLEL_TESTS below
16+ # when changing runner sizes
17+ runs-on :${{ matrix.os == 'macos-latest' && github.repository_owner == 'coder' && 'depot-macos-latest' || matrix.os == 'windows-2022' && github.repository_owner == 'coder' && 'depot-windows-2022-16' || matrix.os }}
1718# This timeout must be greater than the timeout set by `go test` in
1819# `make test-postgres` to ensure we receive a trace of running
1920# goroutines. Setting this to the timeout +5m should work quite well
@@ -31,22 +32,39 @@ jobs:
3132with :
3233egress-policy :audit
3334
35+ # macOS indexes all new files in the background. Our Postgres tests
36+ # create and destroy thousands of databases on disk, and Spotlight
37+ # tries to index all of them, seriously slowing down the tests.
38+ -name :Disable Spotlight Indexing
39+ if :runner.os == 'macOS'
40+ run :|
41+ sudo mdutil -a -i off
42+ sudo mdutil -X /
43+ sudo launchctl bootout system /System/Library/LaunchDaemons/com.apple.metadata.mds.plist
44+
45+ # Set up RAM disks to speed up the rest of the job. This action is in
46+ # a separate repository to allow its use before actions/checkout.
47+ -name :Setup RAM Disks
48+ if :runner.os == 'Windows'
49+ uses :coder/setup-ramdisk-action@79dacfe70c47ad6d6c0dd7f45412368802641439
50+
3451 -name :Checkout
3552uses :actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
3653with :
3754fetch-depth :1
3855
3956 -name :Setup Go
4057uses :./.github/actions/setup-go
58+ with :
59+ # Runners have Go baked-in and Go will automatically
60+ # download the toolchain configured in go.mod, so we don't
61+ # need to reinstall it. It's faster on Windows runners.
62+ use-preinstalled-go :${{ runner.os == 'Windows' }}
63+ use-temp-cache-dirs :${{ runner.os == 'Windows' }}
4164
4265 -name :Setup Terraform
4366uses :./.github/actions/setup-tf
4467
45- # Sets up the ImDisk toolkit for Windows and creates a RAM disk on drive R:.
46- -name :Setup ImDisk
47- if :runner.os == 'Windows'
48- uses :./.github/actions/setup-imdisk
49-
5068 -name :Test with PostgreSQL Database
5169env :
5270POSTGRES_VERSION :" 13"
5573LC_ALL :" en_US.UTF-8"
5674shell :bash
5775run :|
76+ if [ "${{ runner.os }}" == "Windows" ]; then
77+ # Create a temp dir on the R: ramdisk drive for Windows. The default
78+ # C: drive is extremely slow: https://github.com/actions/runner-images/issues/8755
79+ mkdir -p "R:/temp/embedded-pg"
80+ go run scripts/embedded-pg/main.go -path "R:/temp/embedded-pg"
81+ fi
82+ if [ "${{ runner.os }}" == "macOS" ]; then
83+ # Postgres runs faster on a ramdisk on macOS too
84+ mkdir -p /tmp/tmpfs
85+ sudo mount_tmpfs -o noowners -s 8g /tmp/tmpfs
86+ go run scripts/embedded-pg/main.go -path /tmp/tmpfs/embedded-pg
87+ fi
88+
5889 # if macOS, install google-chrome for scaletests
5990 # As another concern, should we really have this kind of external dependency
6091 # requirement on standard CI?
@@ -72,19 +103,29 @@ jobs:
72103 touch ~/.bash_profile && echo "export BASH_SILENCE_DEPRECATION_WARNING=1" >> ~/.bash_profile
73104 fi
74105
106+ # Golang's default for these 2 variables is the number of logical CPUs.
107+ # Our Windows and Linux runners have 16 cores, so they match up there.
108+ NUM_PARALLEL_PACKAGES=16
109+ NUM_PARALLEL_TESTS=16
75110 if [ "${{ runner.os }}" == "Windows" ]; then
76- # Create a temp dir on the R: ramdisk drive for Windows. The default
77- # C: drive is extremely slow: https://github.com/actions/runner-images/issues/8755
78- mkdir -p "R:/temp/embedded-pg"
79- go run scripts/embedded-pg/main.go -path "R:/temp/embedded-pg"
80- else
81- go run scripts/embedded-pg/main.go
111+ # On Windows Postgres chokes up when we have 16x16=256 tests
112+ # running in parallel, and dbtestutil.NewDB starts to take more than
113+ # 10s to complete sometimes causing test timeouts. With 16x8=128 tests
114+ # Postgres tends not to choke.
115+ NUM_PARALLEL_PACKAGES=8
116+ fi
117+ if [ "${{ runner.os }}" == "macOS" ]; then
118+ # Our macOS runners have 8 cores. We leave NUM_PARALLEL_TESTS at 16
119+ # because the tests complete faster and Postgres doesn't choke. It seems
120+ # that macOS's tmpfs is faster than the one on Windows.
121+ NUM_PARALLEL_PACKAGES=8
82122 fi
83123
84- # Reduce test parallelism, mirroring what we do for race tests.
85- # We'd been encountering issues with timing related flakes, and
86- # this seems to help.
87- DB=ci gotestsum --format standard-quiet -- -v -short -count=1 -parallel 4 -p 4 ./...
124+ # We rerun failing tests to counteract flakiness coming from Postgres
125+ # choking on macOS and Windows sometimes.
126+ DB=ci gotestsum --rerun-fails=2 --rerun-fails-max-failures=1000 \
127+ --format standard-quiet --packages "./..." \
128+ -- -v -p $NUM_PARALLEL_PACKAGES -parallel=$NUM_PARALLEL_TESTS -count=1
88129
89130 -name :Upload test stats to Datadog
90131timeout-minutes :1