Bug 1505330

Summary:	No dhcp on virbr0 on s390 in guest
Product:	Red Hat Enterprise Linux 7	Reporter:	Lukáš Doktor <ldoktor>
Component:	libvirt	Assignee:	Libvirt Maintainers <libvirt-maint>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Virtualization Bugs <virt-bugs>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	7.4-Alt	CC:	laine, rbalakri, rjones
Target Milestone:	rc
Target Release:	---
Hardware:	s390x
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2017-11-02 17:53:29 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 Lukáš Doktor 2017-10-23 10:58:09 UTC

Description of problem:
I'm getting no responses in guest when using virbr0 provided by libvirt. The network seems to work well when the IP addr is assigned manually and bridge used by libvirt.

Version-Release number of selected component (if applicable):
Host: RHEL-ALT-7.5-20171017.n.0
Kernel: https://brewweb.engineering.redhat.com/brew/buildinfo?buildID=606293
Qemu: https://brewweb.engineering.redhat.com/brew/buildinfo?buildID=607145

How reproducible:
Always

Steps to Reproduce:
1. /usr/libexec/qemu-kvm -device virtio-scsi-ccw,id=virtio_scsi_ccw0 -drive id=drive_image1,if=none,format=qcow2,file=rhel
7devel-s390x.qcow2 -device scsi-hd,id=image1,drive=drive_image1 -snapshot -m 2G -nographic -nodefaults -chardev socket,
id=serial_id_serial0,path=/tmp/aaa,server,nowait -device sclpconsole,chardev=serial_id_serial0 -device virtio-net-ccw,n
etdev=net0,mac=DE:AD:BE:EF:AA:BB -netdev bridge,br=virbr0,id=net0

Actual results:
Machine boots, network card available, but no ipv4 address available and `dhclient -d eth0` does not show any answers. When IP address manually assigned, host and guest can communicate just well.

Expected results:
Machine should boot and obtain the IP address automatically

Additional info:
I tried creating machine with `virsh` and they also suffered the same issue. I also tried stop/start the network in virt-manager as well as create a new one without any luck.

Comment 2 Lukáš Doktor 2017-10-23 10:59:47 UTC

I forgot to mention `user` network works well.

Comment 3 Laine Stump 2017-10-25 09:06:56 UTC

Last week Rich filed Bug 1503507 against libvirt in Fedora Rawhide, with similar circumstances and symptoms. In particular, he was using qemu:///session (unprivileged libvirt) which uses qemu-bridge-helper to create the tap device and connect it to the bridge, and he also saw failure to acquire an IP with dhclient, but everything worked fine when an IP address was manually assigned.

His suspicion is that it is a host kernel issue, and I don't have an alternative explanation. If there is an older kernel you can try to see if the problem disappears, that would be useful information.

Comment 4 Lukáš Doktor 2017-11-02 17:53:29 UTC

Today I tried the RHEL-ALT-7.4-20171102.n.0-Server-s390x-dvd1.iso and it worked well. Either it's because I used directly LPAR (previously it was z/vm) to run the guest or this was fixed. I'll re-open this if I see this issue again.