Bug 1882490
Summary: | Azure installer misses hyphen in master NIC names | ||
---|---|---|---|
Product: | OpenShift Container Platform | Reporter: | Jim Minter <jminter> |
Component: | Installer | Assignee: | Russell Teague <rteague> |
Installer sub component: | openshift-installer | QA Contact: | To Hung Sze <tsze> |
Status: | CLOSED ERRATA | Docs Contact: | |
Severity: | medium | ||
Priority: | medium | CC: | dcaldwel, dramseur, ffranz, mjudeiki, padillon, tsze, wking |
Version: | 4.4 | Keywords: | Reopened, ServiceDeliveryImpact |
Target Milestone: | --- | ||
Target Release: | 4.9.0 | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: |
Cause: The network interface on master nodes on Azure are missing a hyphen in the interface name which is inconsistent with other platforms.
Consequence: The inconsistency may cause issues in unexpected ways.
Fix: Add a hyphen to the interface name to be consistent with other platforms.
Result: Mater node interfaces are named the same regardless of the platform.
|
Story Points: | --- |
Clone Of: | Environment: | ||
Last Closed: | 2021-10-18 17:28:52 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
Jim Minter
2020-09-24 17:47:18 UTC
I don't really understand how it is important for installer to match the nic name to what the cluster-api names the nics it creates? If a new machine is created by the machine-api, expecting to get the NIC to be part of the LB just because the names are same are previous is IMO incorrect. Someone should take explicit action to add the machine to LB by looking at the NICs of the latest machine and not depend on some _coincidence_ that the name matches the previous one. There is no expectation that this is supposed to work as it is described in the description of the bug. Also it looks like you are depending on machine deletion using machine-api to leave around the NIC since it seems like you are trying to re-use it. I think the machine-api is leaking the NIC here. Brought it up with the team during one the standups and 1. The expected behaviour mentioned in the description is not correct, or not how it's supposed to work. since these LBs are not managed by the cluster, in DR scenarios the user is expected to update the backend with the new machines. so this is not be fixed by the installer team to match the names so that users can _coincidently_ depend on this behaviour. 2. We gave a thought to fixing this for new clusters to maybe allow this to work for ARO, backporting or changing existing clusters was rejected. But as in (1) we decided that changing new clusters has capacity to introduce dependencies that we do not intend to support. So even this was dropped. 3. If the user would the machine-api to add the machine to the LB because I think they have some fields, that could be an RFE 4. But I think the best long term solution would be make these API lbs managed by cluster itself using maybe SLBs are mentioned in https://github.com/openshift/enhancements/pull/459 (3) or (4) should be tracked in a JIRA RFE and not a bug. Can we get some context of why this was reopened? Particularly a response to the first (or second in the case of ARO) point in https://bugzilla.redhat.com/show_bug.cgi?id=1882490#c2 would be helpful. Opened a PR to address code consistency issue. > $ git grep '${var.cluster_id}-master-${count.index}' | wc -l 21 > $ git grep '${var.cluster_id}-master${count.index}' | wc -l 1 4.8.0 tszeaz071421d-t9pvr-master-0 tszeaz071421d-t9pvr-master0-nic I'm resetting severity and priority to medium because an explanation for why this bug is difficult to fix was given in comment 2, then a solution was proposed to align things and that was rejected. I don't want this affecting SLO numbers and taking priority over other bugs which may more readily be addressed. openshift-install 4.9.0-0.nightly-2021-08-18-144658 gpei-shared-0819-08190939-master-0-nic 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 (Moderate: OpenShift Container Platform 4.9.0 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-2021:3759 |