Bug 1762217

Summary:	the conversion pod does not start with blockMode PVCs
Product:	OpenShift Container Platform	Reporter:	Tomas Jelinek <tjelinek>
Component:	Console Kubevirt Plugin	Assignee:	Marek Libra <mlibra>
Status:	CLOSED ERRATA	QA Contact:	Nelly Credi <ncredi>
Severity:	high	Docs Contact:
Priority:	high
Version:	4.2.0	CC:	aos-bugs, cnv-qe-bugs, dagur, mlibra, rhrazdil
Target Milestone:	---
Target Release:	4.3.0
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:	Cause: The kubevirt-storage-class-defaults ConfigMaps setting is not reflected for VMware imports Consequence: blockMode PVC can not be used for VMware VM import Fix: Fixed, the storage class defaults are used when requesting for VMware imported disks Result:	Story Points:	---
Clone Of:	1756328
Clones:	1762218 (view as bug list)		Environment:
Last Closed:	2020-01-23 11:07:48 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	1762218

Description Tomas Jelinek 2019-10-16 08:49:50 UTC

+++ This bug was initially created as a clone of Bug #1756328 +++

How reproducible:
100%

Steps to Reproduce:
1. Have a kubevirt-storage-class-defaults config map which contains "volumeMode: Block" configured for a storage class X
2. Go to the new VM dialog and try to import a VM
3. In the storage section, pick X as a storage class for any or all of the disks
4. Submit the dialog

Actual results:
The conversion pod will never start because it fails to mount the PVCs

Expected results:
The conversion succeeds

Additional info:
The problem is, that if the volume mode is Block, pvc should be mounted using volumeDevices and not volumeMounts

--- Additional comment from Tomáš Golembiovský on 2019-10-10 13:41:59 UTC ---

I have modified entrypoint script to handle the block devices. The change is here:

    https://github.com/oVirt/v2v-conversion-host/commit/cfe8e00f9d6005b6e19aac67be4c50b3d38724fe

It requires corresponding change in the pod definition where the disks should be removed from `volumeMounts` section and included in `volumeDevices` section like this:

      volumeDevices:
        - name: hdd1
          devicePath: /dev/v2v-disk1

The device path is important and must conform to the name /dev/v2v-diskX where X is a number. This is how the entrypoint will find them during execution.

--- Additional comment from Marek Libra on 2019-10-11 12:54:08 UTC ---

UI patch: https://github.com/kubevirt/web-ui-components/pull/557

Configuration of the temporary disk still needs to be resolved, discussion in progress within the PR.

Comment 1 Marek Libra 2019-10-18 07:11:48 UTC

Patch for temporary disk: https://github.com/kubevirt/web-ui-components/pull/561

Comment 3 Radim Hrazdil 2019-10-18 10:22:41 UTC

Verified that with the fix included from https://github.com/kubevirt/web-ui-components/pull/561, the conversion pod disk has correct accessMode RWO and volumeMode Filesystem and successfully binds using rook-ceph storage class.
The VM import started as expected.

Comment 6 errata-xmlrpc 2020-01-23 11:07:48 UTC

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, 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/RHBA-2020:0062