RepoMirrors
/
openssh
mirror of git://anongit.mindrot.org/openssh.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
openssh/.github/workflows/selfhosted.yml

123 lines
4.2 KiB
YAML
Raw Normal View History

Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
name: C/C++ CI self-hosted
on:
push:
Rerun tests on changes to Makefile.in in any dir.
2022-11-03 12:07:50 +00:00
paths: [ '**.c', '**.h', '**.m4', '**.sh', '.github/**', '**/Makefile.in', 'configure.ac' ]
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
jobs:
selfhosted:
Only run selfhosted tests from selfhosted repo.
2021-02-17 07:27:36 +00:00
if: github.repository == 'openssh/openssh-portable-selfhosted'
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
runs-on: ${{ matrix.host }}
Increase selfhosted job timeout. The default job timeout of 360 (6h) is not enough to complete the regress tests for some of the slow VMs depending on the load on the host. Increase to 600 (10h).
2022-11-03 02:41:16 +00:00
timeout-minutes: 600
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
env:
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
HOST: ${{ matrix.host }}
Rename "os" in matrix to "target". This is in preparation to distinguish this from the host that the runner runs on in case where they are separate (eg VMs).
2022-11-26 22:28:20 +00:00
TARGET_HOST: ${{ matrix.target }}
Make "config" in matrix singular and pass in env. This will allow the startup scripts to adapt their behaviour based on the type and config.
2022-11-27 01:14:00 +00:00
TARGET_CONFIG: ${{ matrix.config }}
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
strategy:
fail-fast: false
# We use a matrix in two parts: firstly all of the VMs are tested with the
Rename "os" in matrix to "target". This is in preparation to distinguish this from the host that the runner runs on in case where they are separate (eg VMs).
2022-11-26 22:28:20 +00:00
# default config. "target" corresponds to a label associated with the
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
# worker. The default is an ephemeral VM running under libvirt.
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
matrix:
Rename "os" in matrix to "target". This is in preparation to distinguish this from the host that the runner runs on in case where they are separate (eg VMs).
2022-11-26 22:28:20 +00:00
target:
Add Alpine Linux test VM.
2022-02-22 02:08:59 +00:00
- alpine
Add CentOS 7 test targets.
2023-02-10 03:26:44 +00:00
- centos7
Add debian-i386 test target.
2022-02-26 03:10:41 +00:00
- debian-i386
Add fbsd13 target.
2021-04-16 08:14:25 +00:00
- dfly30
- dfly48
Add dfly60 target.
2021-06-02 01:17:54 +00:00
- dfly60
Add dfly62 test target.
2022-11-11 00:25:37 +00:00
- dfly62
Retire dfly58 test VM. Add dfly64.
2023-07-27 00:29:44 +00:00
- dfly64
Add fbsd10 test target.
2021-04-26 03:35:44 +00:00
- fbsd10
Add fbsd13 target.
2021-04-16 08:14:25 +00:00
- fbsd12
- fbsd13
Add minix3 test target.
2021-05-27 11:13:38 +00:00
- minix3
Add nbsd3, nbsd4 and nbsd9 test targets.
2021-04-25 03:01:34 +00:00
- nbsd3
- nbsd4
Add nbsd8 test target.
2021-04-20 13:31:29 +00:00
- nbsd8
Add nbsd3, nbsd4 and nbsd9 test targets.
2021-04-25 03:01:34 +00:00
- nbsd9
Add obsd51 target.
2021-04-16 23:55:47 +00:00
- obsd51
Add obsd67 test target.
2021-04-23 04:41:13 +00:00
- obsd67
Add obsd69 test target.
2021-05-01 10:07:47 +00:00
- obsd69
Add OpenBSD 7.0 target. Retire 6.8.
2021-12-25 05:42:51 +00:00
- obsd70
Add obsd72 and obsd73 test targets.
2023-08-13 05:01:27 +00:00
- obsd72
- obsd73
Move obsdsnap test VMs to ephemeral runners.
2022-12-13 09:48:56 +00:00
- obsdsnap
- obsdsnap-i386
Add openindiana test target.
2021-04-23 00:28:28 +00:00
- openindiana
Add fbsd13 target.
2021-04-16 08:14:25 +00:00
- sol10
- sol11
Make "config" in matrix singular and pass in env. This will allow the startup scripts to adapt their behaviour based on the type and config.
2022-11-27 01:14:00 +00:00
config:
Rename "os" in matrix to "target". This is in preparation to distinguish this from the host that the runner runs on in case where they are separate (eg VMs).
2022-11-26 22:28:20 +00:00
- default
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
host:
- libvirt
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
include:
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
# Then we include extra libvirt test configs.
- { target: aix51, config: default, host: libvirt }
Add CentOS 7 test targets.
2023-02-10 03:26:44 +00:00
- { target: centos7, config: pam, host: libvirt }
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
- { target: debian-i386, config: pam, host: libvirt }
- { target: dfly30, config: without-openssl, host: libvirt}
- { target: dfly48, config: pam ,host: libvirt }
- { target: dfly58, config: pam, host: libvirt }
- { target: dfly60, config: pam, host: libvirt }
- { target: dfly62, config: pam, host: libvirt }
- { target: fbsd10, config: pam, host: libvirt }
- { target: fbsd12, config: pam, host: libvirt }
- { target: fbsd13, config: pam, host: libvirt }
- { target: nbsd8, config: pam, host: libvirt }
- { target: nbsd9, config: pam, host: libvirt }
- { target: openindiana, config: pam, host: libvirt }
- { target: sol10, config: pam, host: libvirt }
- { target: sol11, config: pam-krb5, host: libvirt }
- { target: sol11, config: sol64, host: libvirt }
# VMs with persistent disks that have their own runner.
- { target: win10, config: default, host: win10 }
- { target: win10, config: cygwin-release, host: win10 }
# Physical hosts, with either native runners or remote via ssh.
- { target: ARM, config: default, host: ARM }
- { target: ARM64, config: default, host: ARM64 }
- { target: ARM64, config: pam, host: ARM64 }
- { target: debian-riscv64, config: default, host: debian-riscv64 }
obsd-arm64 host is real hardware... so put in the correct config location.
2023-08-21 09:47:58 +00:00
- { target: obsd-arm64, config: default, host: obsd-arm64 }
Rework how selfhosted tests interact with runners. Previously there was one runner per test target (mostly VMs). This had a few limitations: - multiple tests that ran on the same target (eg multiple build configs) were serialized on availability or that runner. - it needed manual balancing of VMs over host machines. To address this, make VMs that use ephemeral disks (ie most of them) all use a pool of runners with the "libvirt" label. This requires that we distinguish between "host" and "target" for those. Native runners and VMs with persistent disks (eg the constantly-updated snapshot ones) specify the same host and target. This should improve test throughput.
2022-11-27 04:26:22 +00:00
- { target: openwrt-mips, config: default, host: openwrt-mips }
- { target: openwrt-mipsel, config: default, host: openwrt-mipsel }
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
steps:
Move vmshutdown to first step. If a previous run on a physical runner has failed to clean up, the next run will fail because it'll try to check out the code to a broken directory mount. Make cleanup the first step.
2022-07-15 11:31:48 +00:00
- name: shutdown VM if running
run: vmshutdown
Run vm startup and shutdown from runner temp dir. Should work even if the github workspace dir is on a stale sshfs mount.
2022-11-08 00:33:25 +00:00
working-directory: ${{ runner.temp }}
Update checkout and upload actions. Update actions/checkout and actions/upload-artifact to main branch for compatibility with node.js v16.
2022-11-07 22:21:10 +00:00
- uses: actions/checkout@main
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
- name: autoreconf
run: autoreconf
- name: startup VM
run: vmstartup
Run vmstartup from temp dir. This will allow us to create ephemeral disk images per-runner.
2022-11-27 01:19:37 +00:00
working-directory: ${{ runner.temp }}
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
- name: configure
Make "config" in matrix singular and pass in env. This will allow the startup scripts to adapt their behaviour based on the type and config.
2022-11-27 01:14:00 +00:00
run: vmrun ./.github/configure.sh ${{ matrix.config }}
Always save config.h as build artifact. Should allow better comparison between failing and succeeding test platforms.
2021-12-25 23:31:15 +00:00
- name: save config
Update checkout and upload actions. Update actions/checkout and actions/upload-artifact to main branch for compatibility with node.js v16.
2022-11-07 22:21:10 +00:00
uses: actions/upload-artifact@main
Always save config.h as build artifact. Should allow better comparison between failing and succeeding test platforms.
2021-12-25 23:31:15 +00:00
with:
Make "config" in matrix singular and pass in env. This will allow the startup scripts to adapt their behaviour based on the type and config.
2022-11-27 01:14:00 +00:00
name: ${{ matrix.target }}-${{ matrix.config }}-config
Always save config.h as build artifact. Should allow better comparison between failing and succeeding test platforms.
2021-12-25 23:31:15 +00:00
path: config.h
Add make clean step to tests.
2021-09-29 03:28:56 +00:00
- name: make clean
run: vmrun make clean
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
- name: make
run: vmrun make
- name: make tests
Make "config" in matrix singular and pass in env. This will allow the startup scripts to adapt their behaviour based on the type and config.
2022-11-27 01:14:00 +00:00
run: vmrun ./.github/run_test.sh ${{ matrix.config }}
Increase test timeout to allow slow VMs to finish
2022-04-01 12:21:06 +00:00
timeout-minutes: 600
Save config.h and config.log on failure too.
2021-04-03 05:59:10 +00:00
- name: save logs
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
if: failure()
Update checkout and upload actions. Update actions/checkout and actions/upload-artifact to main branch for compatibility with node.js v16.
2022-11-07 22:21:10 +00:00
uses: actions/upload-artifact@main
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
with:
Make "config" in matrix singular and pass in env. This will allow the startup scripts to adapt their behaviour based on the type and config.
2022-11-27 01:14:00 +00:00
name: ${{ matrix.target }}-${{ matrix.config }}-logs
Upload valgrind logs on failure.
2021-02-20 02:06:25 +00:00
path: |
Save config.h and config.log on failure too.
2021-04-03 05:59:10 +00:00
config.h
config.log
Upload valgrind logs on failure.
2021-02-20 02:06:25 +00:00
regress/*.log
Adjust test jobs for new log directory.
2023-03-01 12:23:02 +00:00
regress/log/*
Upload valgrind logs on failure.
2021-02-20 02:06:25 +00:00
regress/valgrind-out/
Add self-hosted runners for VMs of other platforms. Github only hosts a limited number of platforms, and the runner code is only supported on slightly wider range of platforms. To increase our test coverage beyond that, we run the runner natively on a VM host, where it runs a jobs that boot VMs of other platforms, waits for them to come up then runs the build and test by ssh'ing into the guest. This means that the minimum dependencies for the guests are quite low (basically just sshd, a compiler and make). The interface to the VM host is fairly simple (basically 3 scripts: vmstartup, vmrun and vmshutdown), but those are specific to the VM host so are not in the public repo. We also mount the working directory on the host via sshfs, so things like artifact upload by the runner also work. As part of this we are moving the per-test-target configs into a single place (.github/configs) where there will be referenced by a single short "config" key. I plan to make the github-hosted runners use this too. The self-hosted runners are run off a private repo on github since that prevents third parties from accessing them[0], and since runner quota is limited on private repos, we avoid running the tests we run on the public repo. [0] https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#self-hosted-runner-security-with-public-repositories
2021-01-15 03:11:43 +00:00
- name: shutdown VM
if: always()
run: vmshutdown
Run vm startup and shutdown from runner temp dir. Should work even if the github workspace dir is on a stale sshfs mount.
2022-11-08 00:33:25 +00:00
working-directory: ${{ runner.temp }}