report and record the actual booted kernel with test results
Bug #1491865 reported by
Andy Whitcroft
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
autopkgtest (Ubuntu) |
Fix Released
|
High
|
Martin Pitt |
Bug Description
As we often have 2 or more kernels installed when testing the package list is not sufficient to tell us which kernel we intended to run. Nor indeed does having the intended kernel installed tell us we actually running it after the reboot. It would be nice to:
1) report the kernel version uname -rv style as the first action following a reboot, allowing us to confirm the actual version booted each time.
2) perhaps report the kernel version and other information in the results.tar, say in a sysinfo.yaml.
If we chose a .yaml for this info we could expand it later to include filesystem info for example.
Related branches
Changed in autopkgtest (Ubuntu): | |
status: | New → In Progress |
importance: | Undecided → High |
assignee: | nobody → Martin Pitt (pitti) |
Changed in autopkgtest (Ubuntu): | |
milestone: | none → ubuntu-15.09 |
To post a comment you must log in.
As discussed on IRC, we want something like this for 2:
>>> print(json. dumps({ 'kernel_ version' : '1.2', 'test_kernel_ versions' : [('test2', 'reboot_ to_backport' , '1.3~utopic')]}, indent=2)) kernel_ versions" : [ reboot_ to_backport" ,
{
"kernel_version": "1.2",
"test_
[
"test2",
"
"1.3~utopic"
]
]
}
kernel_version is the version right after the initial testbed setup/dist- upgrade/ reboot. Any test which triggers a reboot which ends up in a kernel != kernel_version will get an entry in that test_kernel_ versions list. Any test/reboot which is not there can be assumed to use kernel_version. IOW, for pretty much all expected cases test_kernel_ versions should be completely absent.
We use json instead of yaml as the latter requires an external Python module.