[release-4.6] templates: use Afterburn for setting GCP hostnames#2286
Conversation
The only platform where over-log hostnames have been encountered is on GCP. The code has proven buggy, racy and caused a bunch of BZ. To unwind this mess, on GCP, the new behavior: - leaves disabling NetworkManager on GCP - on each boot run Afterburn to fetch the hostname and writes it an ephemeral location - uses the existing checks to truncate the length Finally, this DROPS the NetworkManager dispatcher. FCOS/RHCOS is pursuing a more permanent solution. Signed-off-by: Ben Howard <ben.howard@redhat.com>
Co-authored-by: Vadim Rutkovsky <roignac@gmail.com>
openshift-ci-robot
requested review from
kikisdeliveryservice and
sinnykumari
|
@LorbusChris @darkmuggle If you want this backported to 4.6, you'll need to make a bz for this and for the original 4.7 bz so it can get cherrypick approval. |
|
Clean cherrypick and is needed by OKD to set hostname on GCP /approve |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: openshift-cherrypick-robot, sinnykumari The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci-robot
added
the
approved
|
/test e2e-gcp-op |
|
@openshift-cherrypick-robot: The following tests failed, say
Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
In regards to #2286 (comment) Given the extreme caution that I have taken on the original PR and on the VSphere bug, I want to slow-roll the backport until at least 4.7 is done soaking. This change introduces a behavior change by dropping the NM dispatcher hook. The bug-laden wasteland of hostname issues tells me that caution is warranted. /hold |
openshift-ci-robot
added
the
do-not-merge/hold
|
@darkmuggle SGTM 👍 |
|
Systemd has reverted the fedora default hostname https://src.fedoraproject.org/rpms/systemd/pull-request/42 which begs the question of whether we should backport this; the motivation for this PR is OKD. Unless there is a clear need to change the behavior of GCP hostname handling on OCP, I am inclined to wait until there's a clear need. |
|
agreed, with the systemd revert this isn't required anymore. I'll close this PR |
|
@LorbusChris: Closed this PR. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
7 participants
This is an automated cherry-pick of #2217
/assign LorbusChris