Description of problem ---------------------- Some workflows / use cases don't work anymore because /etc/redhat-release seems to be a broken symlink. Version of all relevant components ---------------------------------- OCP Version: 4.11.0-0.nightly-2022-03-20-160505 RHCOS Version: 411.85.202203181601-0 Steps to reproduce ------------------ 1. Deploy an OpenShift 4.11 cluster 2. Log-in to one node and check /etc/redhat-release. Actual results -------------- 1. /etc/redhat-release is a symlink pointing to a non-existent file (/usr/lib/system-release) 2. /etc/redhat-release is an absolute symlink Expected results ---------------- 1. /etc/redhat-release is a symlink pointing to a valid file (/usr/lib/redhat-release is used in RHCOS-4.10) 2. /etc/redhat-release is a relative symlink (../usr/lib/redhat-release is used in RHCOS-4.10) Additional info --------------- This breaks one script used for SR-IOV network operator configuration (https://github.com/openshift/sriov-network-operator/blob/master/bindata/scripts/enable-kargs.sh) which expects /etc/redhat-release to exist and to be a relative symlink because it is accessed through /host mountpoint.
This is from https://github.com/openshift/os/pull/713 > 1. Deploy an OpenShift 4.11 cluster BTW, as of lately you can just do e.g.: $ podman run --rm -ti registry.ci.openshift.org/rhcos-devel/rhel-coreos:4.11 (This is part of https://github.com/openshift/enhancements/pull/1032 , we will be moving the image to a more official place) > 1. /etc/redhat-release is a symlink pointing to a non-existent file (/usr/lib/system-release) Yeah, definitely a bug. (But that said, why are you parsing redhat-release? The new cross-distribution standard is `/usr/lib/os-release` (or previously `/etc/os-release`))
(In reply to Colin Walters from comment #1) > (But that said, why are you parsing redhat-release? The new > cross-distribution standard is `/usr/lib/os-release` (or previously > `/etc/os-release`)) I am innocent :). I'm just a user of the SR-IOV operator. This script was written by the OpenShift Network SR-IOV team. I don't know if there is a real motivation behind it or they just used /etc/redhat-release out of habit.
We will fix that but the script should be looking for /run/ostree-booted instead of greping a release file. Will file an issue.
Filed: https://github.com/openshift/sriov-network-operator/issues/647
Verify passed with rhcos-411.86.202206062029-0-qemu.x86_64.qcow2 [core@cosa-devsh ~]$ ll /etc/redhat-release lrwxrwxrwx. 1 root root 23 Jun 6 20:35 /etc/redhat-release -> /usr/lib/system-release [core@cosa-devsh ~]$ ll /usr/lib/system-release -rw-r--r--. 2 root root 45 Jan 1 1970 /usr/lib/system-release
Remove the bootimage tracker bug as it does not need
Verify passed with latest image rhcos-411.86.202206090101-0-qemu.x86_64.qcow2 [core@cosa-devsh ~]$ ll /etc/redhat-release lrwxrwxrwx. 1 root root 23 Jun 9 01:07 /etc/redhat-release -> /usr/lib/system-release [core@cosa-devsh ~]$ ll /usr/lib/system-release -rw-r--r--. 2 root root 45 Jan 1 1970 /usr/lib/system-release
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: OpenShift Container Platform 4.11.0 bug fix and security update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2022:5069