kexec-tools/tests
Tao Liu f4ab396574 selftest: run-test.sh: wait for subprocess instead of kill it
When run tests with 2 VMs, for example nfs/ssh kdump tests, client VM will do the
crash and dump, server VM will do vmcore saving and if-vmcore-exists
check.

Previously, when client VM finishes running, run-test.sh will kill the lead background
process, and then check if server VM has outputted "TEST PASSED" or "TEST FAILED" string.
However it didn't wait for server VM to finish. As a result, the server VM's final
outputs are not collected and checked, leaving the test result as "TEST RESULT NOT FOUND"
sometimes.

For example, the following is the pstree status of $(jobs -p) before it
gets killed. We can see the server VM is still running:

run-test.sh,172455 /root/kexec-tools/tests/scripts/run-test.sh --console nfs-early-kdump
  └─run-test.sh,172457 /root/kexec-tools/tests/scripts/run-test.sh --console...
      └─timeout,172480 --foreground 10m /root/kexec-tools/tests/scripts/run-qemu...
          └─qemu-system-x86,172481 -enable-kvm -cpu host -nodefaults...
              ├─{qemu-system-x86},172489
              ├─{qemu-system-x86},172492
              ├─{qemu-system-x86},172493
              ├─{qemu-system-x86},172628
              └─{qemu-system-x86},172629

In this patch, we will wait for $(jobs -p) to finish, in order to get
the complete output of test results.

Signed-off-by: Tao Liu <ltao@redhat.com>
Acked-by: Coiby Xu <coxu@redhat.com>
2022-01-24 11:17:52 +08:00
..
scripts selftest: run-test.sh: wait for subprocess instead of kill it 2022-01-24 11:17:52 +08:00
Makefile selftest: Make test_base_image depends on EXTRA_RPMS 2021-06-08 22:21:47 +08:00
README selftest: Add document for selftests 2020-09-17 10:42:58 +08:00

=====================
Kexec Kdump Self-test
=====================


Introduction
============
The self-tests here are useful for quick sanity tests for new patches, and also helpful for debugging issues.


How it works
============
All tests are run within VMs using qemu. By default, VM images are based on Fedora Cloud image, and the image for each test run is a layered qcow2 snapshot on top of the base image.
Test images are managed by Makefile, so if there are any code change in the kexec-tools repository, `make` command will detect that and only rebuild the top image layer. This makes the test runs boot fast and each test run is clean.


Basic usage
===========
Before you start, you can make the self-tests use your own base image by running following command:

`make clean && make BASE_IMAGE=<path/to/your/image>`

This is helpful if you have a slow network, else self-test will try to download the cloud image from Fedora's official website using `wget`.

- Use the following command to run all tests:
  $ make test-run

  All available tests will be executed.

  Test artifacts are stored in output/<testcase>

- For easier debugging, you can run only on test with the following command:
  $ make TEST_CASE=<testcase> test-run

  This way, VM's console is directly connected to stdin/out so debugging will be easier.
  If there are multiple VMs used in a test case, the VM performing actual kdump/kexec operation will be connected to stdin/out.

Test Cases
==========
Each test case is a folder under scripts/testcases/, a test case folder will contain at least one executable shell script, and each script should contain two functions: "on_build" and "on_test".

"on_build" is called when building the test image, which can instruct the self-test framework to install packages or create files, etc.
"on_test" is called when VM finished booting, which can get the boot count by calling "get_test_boot_count" and determine what to do. It should call "test_passed" on success, and call "test_failed" on failure. "test_aborted" is called when unexpected behavior occurs.

When there are multiple scripts in a single test case folder, they will spawn VMs in lexical order, and the last VM is considered the VM performing the actual test. Other VMs could be hosting test required service. This is useful for the network dump test. However, "test_passed" or "test_failed" or "test_aborted" could be called in any of these VMs, so during network kdump test, the dump target can also terminate the test and mark it passed when a valid vmcore is detected.


Debugging
=========

- When the test VM boots, you can append "no_test" to kernel args in grub, which tells the test services to quit early.

- You can launch the VMs manually or inspect the image after ran a test.

  Test images are located as:

    output/<testcase>/<vm-name>.img

  Test images' corresponding qemu command are located as:

    output/<testcase>/<vm-name>.qemu_cmd

  To repeat/debug a test manually, you should launch all VMs in output/<testcase> menu in lexical order, and append 'no_test' in the last VM's grub cmdline, then VM will hang on login prompt, login with root/fedora. Test script is located as /kexec-kdump-test/test.sh

- If you just want to inspect the images file content, you can also use scripts/spawn-image-shell.sh <test-image> to spawn a shell in the image quickly.