Deploy Nova libvirt with Kolla kubernetes get error for libvirt
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| | kolla |
Undecided
|
Unassigned | ||
Bug Description
I've already change the nova-libvirt yaml file, comment the "rpm -e qemu-system-x86 || true;"
=======
linux@kolla-
diff --git a/helm/
index 58caaaf..e851cf0 100644
--- a/helm/
+++ b/helm/
@@ -91,8 +91,12 @@ spec:
# Following is added to workaround bug
# https:/
+{{- if $localVals.
rpm -e qemu-system-x86 || true;
+{{- end }}
+{{- if $localVals.
cp -a {{ $containerConfi
+{{- end }}
=======
Use "kubectl logs nova-compute** -n kolla" I can see the libvirtd starts.
But check the libvirtd I find the deployment failed
still got the error as below:
Image is 5.0.1, Pike,
Libvirt version is 3.6.0
Steps to Reproduce: Minimized, easy-to-follow steps that will trigger the bug. Include any special setup steps.
1.helm install --debug kolla-kubernete
2. wait and get the logs from nova-libvirt container
3. docker inspect <nova libvirt container id>
4. Get the logs directory from host()
{
},
5. Check the libvirtd logs
...
Container image:
haerwu/
Actual Results: What happened after performing the above steps.
Check the logs from the container.
2017-11-17 05:57:57.385+0000: 18161: info : libvirt version: 3.6.0, package: 1~bpo9+1 (Riku Voipio <email address hidden> Mon, 14 Aug 2017 14:02:02 +0300)
2017-11-17 05:57:57.385+0000: 18161: info : hostname: kolla-kubernete
2017-11-17 05:57:57.385+0000: 18161: error : daemonSetupNetw
server certificate path set to match server key
| description: | updated |
| Kevin Zhao (kevin-zhao) wrote : | #1 |
| Rich Wellum (richwellum) wrote : | #2 |
I also see this in stable/pike, running kolla-kubernetes.
| Oliver Walsh (owalsh) wrote : | #4 |
Looks like libvirt TLS is enabled (listen_tls=1 in libvirtd.conf) but certificates have not been setup.
ca_file="" so I assume no attempt is being made to configure TLS. Setting listen_tls=0 will probably get this working.
However setting listen_tcp=1, auth_tcp="none", listen_addr = "0.0.0.0" is a major security issue. Anybody on the network can virsh! This was a CVE in tripleo/
| Rich Wellum (richwellum) wrote : | #5 |
Some debug.
I added a sleep to nova-libvirt container so I could exec in and run kolla_start manually.
()[root@ubunturich /]# kolla_start
INFO:__
INFO:__
INFO:__main__:Kolla config strategy set to: COPY_ONCE
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
Running command: '/usr/sbin/libvirtd --listen'
()[root@ubunturich /]# cat /var/log/
2017-11-28 04:45:14.274+0000: 9203: info : libvirt version: 3.2.0, package: 14.el7_4.3 (CentOS BuildSystem <http://
2017-11-28 04:45:14.274+0000: 9203: info : hostname: ubunturich
2017-11-28 04:45:14.274+0000: 9203: error : daemonSetupNetw
()[root@ubunturich /]#
| Oliver Walsh (owalsh) wrote : | #6 |
Doh, I was looking at the test env in comment #4.
Could you include the contents of /etc/libvirtd.conf as it definitely seems related to the TLS config.
| Jeffrey Zhang (jeffrey4l) wrote : | #7 |
This is fixed by https:/
| Kevin Zhao (kevin-zhao) wrote : | #8 |
Hi Jeffery, Thanks for reminding
I will port this fix to kolla-kubernetes to verify


If not apply the patch in the above:
error log: kolla/config_ files/secrets' : No such file or directory main__: Loading config file at /var/lib/ kolla/config_ files/config. json main__: Validating config file main__: Copying service configuration files main__: Deleting /etc/libvirt/ libvirtd. conf main__: Copying /var/lib/ kolla/config_ files/libvirtd. conf to /etc/libvirt/ libvirtd. conf main__: Setting permission for /etc/libvirt/ libvirtd. conf main__: Deleting /etc/libvirt/ qemu.conf main__: Copying /var/lib/ kolla/config_ files/qemu. conf to /etc/libvirt/ qemu.conf main__: Setting permission for /etc/libvirt/ qemu.conf main__: Writing out command to execute
/bin/bash: line 3: rpm: command not found
cp: cannot stat '/var/lib/
INFO:__
INFO:__
INFO:__main__:Kolla config strategy set to: COPY_ONCE
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
Running command: '/usr/sbin/libvirtd --listen'
If applied and re deploy: main__: Loading config file at /var/lib/ kolla/config_ files/config. json main__: Validating config file main__: Copying service configuration files main__: Deleting /etc/libvirt/ libvirtd. conf main__: Copying /var/lib/ kolla/config_ files/libvirtd. conf to /etc/libvirt/ libvirtd. conf main__: Setting permission for /etc/libvirt/ libvirtd. conf main__: Deleting /etc/libvirt/ qemu.conf main__: Copying /var/lib/ kolla/config_ files/qemu. conf to /etc/libvirt/ qemu.conf main__: Setting permission for /etc/libvirt/ qemu.conf main__: Writing out command to execute
INFO:__
INFO:__
INFO:__main__:Kolla config strategy set to: COPY_ONCE
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
INFO:__
Running command: '/usr/sbin/libvirtd --listen'
But libvirtd error is always there