tests: add kdump over NFS #3922
Conversation
|
/hold |
jbtrystram
changed the title
|
/retest |
jbtrystram
force-pushed
the
nfs-kdump-test
branch
from
3750b7d to
1746052
Compare
jbtrystram
removed
do-not-merge/hold
hold-for-external
|
ah, I got excited because I saw dracut 105 got in stable but I forgot we pinned the previous version because it broke multipath tests... still waiting I guess |
|
/retest |
This way we have good coverage of most-used kdump features. Some context on the NFS kdump configuration: coreos/fedora-coreos-tracker#1729 This was previously merged in [1] then reverted in [2] because the nfs server container was not multi-arch, causing the pipeline to trip on it. It's also not functionning on systemd256 (so anything f41 and above), see [3] This requires coreos#3917 for the multi-arch container, and coreos#3921 [1] coreos@b10d8dc [2] coreos@af1468c [3] rhkdump/kdump-utils#52
jbtrystram
force-pushed
the
nfs-kdump-test
branch
from
1746052 to
683600f
Compare
|
rebased. The test pass on a rawhide build which have dracut 105, so we just need to wait for testing-devel to rebase to F42 content. |
This should happen next week so we won't have to wait long. |
|
[2025-04-17T07:43:54.376Z] --- PASS: kdump.crash.nfs (140.24s) 🎉 |
|
Can we get the commit message updated to be more accurate now? i.e. it states that it doesn't work on anything f41 and above, which clearly isn't accurate now because we are on f42 and it passes. |
| }) | ||
| } | ||
|
|
||
| // This function test the remote kdump feature by: |
There was a problem hiding this comment.
| // This function test the remote kdump feature by: | |
| // This function tests the remote kdump feature by: |
| } | ||
|
|
||
| // crash the kernel | ||
| // use systemd-run because direclty calling `echo c > ...` will always |
There was a problem hiding this comment.
| // use systemd-run because direclty calling `echo c > ...` will always | |
| // use systemd-run because directly calling `echo c > ...` will always |
This was previously merged in [1] then reverted in [2] because the nfs
server container was not multi-arch, causing the pipeline to trip on it.
It's also not functionning on systemd256 (so anything f41 and above),
see [3]
This requires #3917 for
the multi-arch container, and
#3921
[1] b10d8dc
[2] af1468c
[3] rhkdump/kdump-utils#52