Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 2042378

Summary: SNO bootstrap: openshift-etcd fails to start due to missing /etc/resolv.conf
Product: OpenShift Container Platform Reporter: Sagi Dayan <sdayan>
Component: assisted-installerAssignee: Mat Kowalski <mko>
assisted-installer sub component: Installer QA Contact: Udi Kalifon <ukalifon>
Status: CLOSED DUPLICATE Docs Contact:
Severity: unspecified    
Priority: unspecified CC: aos-bugs, bpickard, mfilanov, mko, wking, yliu1, yobshans
Version: 4.9   
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2022-04-20 09:04:04 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Sagi Dayan 2022-01-19 11:25:49 UTC
Description of problem:

bootkube etcd instance is not be able to start due to missing resolv.conf file.

Journal logs indicates: 

Failed to generate sandbox config for pod" err="open /etc/resolv.conf: no such file or directory" pod="openshift-etcd/etcd-bootstrap-member-random-hostname-47bf49cb-4aaa-4eed-8c1d-6ad860d5642b 


Version-Release number of selected component (if applicable):
We found this to happen on both 4.8 and 4.9
(See links)

How reproducible:
Unknown at this point. These cases are reported by assisted-installer. Both cases are SNO's and using SDN Networking.


Actual results:
bootkube etcd fails to start since it has a mount to /etc/resolv.conf that does not exists

Expected results:
resolv.conf should exist and etcdd should run

Additional info:

Comment 1 Ben Nemec 2022-01-20 18:58:39 UTC
I'm going to move this to the assisted installer component because this appears to be a host configuration issue. There doesn't appear to be any SDN involvement and the IPI resolv-prepender script is not used here either.

I'm not sure whether resolv.conf was never present or if it's being removed as part of the deployment process, and I'm not familiar enough with the assisted flow to know what might be causing this. We're going to need to work with the deployer to figure out what's going on here.

Comment 2 Mat Kowalski 2022-01-21 11:50:34 UTC
Looks extremely similar to https://bugzilla.redhat.com/show_bug.cgi?id=1968634 (apart from the fact that in here the empty /etc/resolv.conf appears at the beginning of the SNO installation). Lots of pointers to https://github.com/openshift/assisted-service/blob/v2.0.10/internal/network/manifests_generator.go#L93-L107, but nothing obvious or clear

Comment 5 Michael Filanov 2022-04-20 09:04:04 UTC
Closing as duplicate, it will happen again we can open a jira bug

*** This bug has been marked as a duplicate of bug 1968634 ***