tap device down after nova reboot
Bug #1662131 reported by
Slobodan Blatnjak
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Juniper Openstack |
New
|
Undecided
|
Kumar Harsh | ||
R2.21.x |
New
|
High
|
Kumar Harsh |
Bug Description
Contrail: 2.21
Doing a (soft) reboot on the nova api will cause the tap device to be down.
Centos 7.3
Libvirt 2.0.0-10
qemu-kvm-
contrail 2.21
Another way to test without OpenStack api:
virsh shutdown $instance
virsh start $instance
On a centos 7.2 hypervisor this will put the tap device in an up state.
However with centos7.3 the device will keep in a down state.
Not sure this is a contrail or libvirt/qemu bug.
For now they "fixed" it in the nova code by making it do a hard reboot after the shutdown which will put it back in a good state.
However it would be nice to make it work again with default upstream code.
tags: | added: ebay |
tags: | added: vrouter |
tags: | added: nova |
description: | updated |
information type: | Proprietary → Public |
Changed in juniperopenstack: | |
assignee: | nobody → Srinivasan Venkatakrishnan (srinivn) |
Changed in juniperopenstack: | |
assignee: | Srinivasan Venkatakrishnan (srinivn) → Kumar Harsh (hkumar) |
Changed in juniperopenstack: | |
milestone: | r3.2.10.0 → none |
To post a comment you must log in.
The tap interface is down during a softreboot whereas it is up during an hardreboot.
Contrail 3.2.3
Red Hat 7.3
Libvirt version 2.0.0 release 10.el7_3.9
qemu version 1.5.3 release 126.el7_3.10
1. src_mac address of the VM
### ifconfig -a
eth0 Link encap:Ethernet HWaddr 02:B7:A8:71:AA:14
2. tap interface name of the VM
tapb7a871aa-14 (compute0)
3. tcpdump from tap interface of the VM2 (with softreboot and hardreboot)
###softreboot: openstack server reboot --soft test_power_ management compute0 ~]$ sudo tcpdump -n -i tapb7a871aa-14
#results on the compute node:
[nie@pop199a-
tcpdump: tapb7a871aa-14: That device is not up