[rhos-list] woes with RHOS + Quantum + OVS
Daniel Dumitriu
daniel at dumdan.com
Sat Nov 10 08:06:52 UTC 2012
Sorry, I cannot help with a suggestion.
Your post, however, is very interesting because you brought up Quantum.
Did you use the packages from "epel-testing" ? Did ou install the DHCP
agent, as well ?
How does Nova cooperate with Quantum ? Isn't there a conflict between
"dnsmaq" and "quantum-dhcp" ?
I ran, more or less, into the same problems and I am determined to -
somehow - solve them. Frankly, the virtual network component is crucial
for building "real world configurations" in the cloud.
I do have an instance of libvirt running with OVS, (not on RHEL but on
SL 6.3, because this all happens on my personal systems, home, and I
haven't got that many RHEL private subscriptions) and all my working
environments are running on top of it.
That said, I must add that it is a total hack - I had to force libvirt
to work with OVS, literally.
I think that getting Quantum to work properly in OpenStack, with
libvirt/KVM should be the first development priority, if we are to
provision really isolated, multi-tenant, multi-layer configurations.
(Of course, the second would be to stop naming the VMs "domains", in
libvirt :-) ).
--
___________
Daniel Dumitriu
daniel at dumdan.com
Telephone: +1-416-626-9345
Mobile: +1-416-318-2487
-----Original Message-----
From: Jonathan Mills <jonmills at renci.org>
To: rhos-list at redhat.com
Subject: [rhos-list] woes with RHOS + Quantum + OVS
Date: Mon, 10 Sep 2012 14:50:51 -0400
I have RHOS configured to run with Quantuam and Openvswitch. (I created
my Openvswitch RPMS for RHEL6 by doing a 'git clone', preparing the
source tarball, and using their included spec file.) As far as I can
tell, OVS is happy. Running 'ovs-vsctl show' gives me reasonable output:
# ovs-vsctl show
0c960ad6-2de7-4651-bfa9-9f0b3b134755
Bridge br-int
Port br-int
Interface br-int
type: internal
Port "gw-c83a1b4a-d7"
tag: 4095
Interface "gw-c83a1b4a-d7"
type: internal
Port "eth1"
Interface "eth1"
ovs_version: "1.8.90"
Openstack itself is happy, with all smiley faces.
My problem emerges when I actually launch a VM with 'nova boot'.
$ nova list
+--------------------------------------+--------+--------+----------+
| ID | Name | Status | Networks |
+--------------------------------------+--------+--------+----------+
| 2e446148-9d3f-430a-a4df-7d0ec0329e93 | jbmVM1 | ACTIVE | |
+--------------------------------------+--------+--------+----------+
As you can see above, the 'Networks' column is empty. The result, for
the VM, looks like this (from console log):
SIOCSIFADDR: No such device
eth0: ERROR while getting interface flags: No such device
eth0: ERROR while getting interface flags: No such device
Bind socket to interface: No such device
Failed to bring up eth0.
An interesting piece of this puzzle.....when I read the libvirt XML
description file for this VM, it is literally missing a network device.
Why would that happen?
Other notes:
The way my system is set up, it's a one-node test cluster. So
nova-network and nova-compute run on the same box. This box has an eth0
with a public IP address. It has an eth1 with a 10.x.x.x address. I'm
running my bridge on eth1. I have Openstack configured to have all its
"fixed_ips" in the 10.x.x.x range, and it's "floating_ips" as real
public IPs that should be added onto eth0.
Can anyone think of some items that I can double-check to figure out
what I've done wrong?
email message attachment
-------- Forwarded Message --------
From: Xu, HongnaX <hongnax.xu at intel.com>
To: rhos-list at redhat.com <rhos-list at redhat.com>
Subject: [rhos-list] Some issues about install packages of Openstcak
Date: Tue, 11 Sep 2012 06:48:34 +0000
Hi all
It’s me again. I encountered some issues with openstack preview on RH
enterprise server 6.3.
1. install RHEL.
2. Add myproxy in ~/.bashrc.
3. Wget http://google.com to verify the proxy is working.
4. Subscribe Openstack, I got no other channels except “openstack”,
it is not stabilized, yesterday I can list “Red Hat Enterprise Linux
Server”, I got it by the command “subscription-manager list --available
--proxy=MYPROCY:PORT”,
Question 1: why there is only “Openstack” channel?How could I subscribe
Red Hat Enterprise Linux Server channel ?
5. with the default repos which were installed by the image, I
think the packages with openstack should come from the “Openstack” or
“Red Hat Enterprise Linux Server” channel, am I right? Then I yum search
of yum update, the logs are:
***************************
# yum update
Loaded plugins: product-id, security, subscription-manager
Updating certificate-based repositories.
[Errno 110] Connection timed out
Setting up Update Process
No Packages marked for Update
******************************
#Yum search git
Loaded plugins: product-id, security, subscription-manager
Updating certificate-based repositories.
[Errno 110] Connection timed out
================================== Matched: git
===================================
cdparanoia-libs.x86_64 : Libraries for libcdda_paranoia (Paranoia III).
eggdbus.x86_64 : Experimental D-Bus bindings for GObject
gnupg2.x86_64 : Utility for secure communication and data storage
grubby.x86_64 : Command line tool for updating bootloader configs
hwdata.noarch : Hardware identification and configuration data
libaio.x86_64 : Linux-native asynchronous I/O access library
mailcap.noarch : Helper application and MIME type associations for file
types
mcelog.x86_64 : Tool to translate x86-64 CPU Machine Check Exception
data.
numad.x86_64 : NUMA user daemon
pixman.x86_64 : Pixel manipulation library
python-ethtool.x86_64 : Ethernet settings python bindings
python-gudev.x86_64 : Python (PyGObject) bindings to the GUDev library
python-iwlib.x86_64 : Wireless settings python bindings
python-netaddr.noarch : Pythonic manipulation of IPv4, IPv6, CIDR, EUI
and MAC
: network addresses
setuptool.x86_64 : A text mode system configuration tool
**********************
All the packages listed are not related with git, I don’t why they could
be listed. And why I encounter 110 Error?
Question 2: is it the exactly the reason that I can’t get the packages
of Openstack ?? and how to fix it?
6. With a second machine I subscribe “openstack” “Red Hat Beta”
“Red Hat Enterprise Linux Server”, and set centos repo
in /etc/yum.repos.d, and I could find the packages of dnsmasq-utils and
git, Though encountered 110 Error with yum command either.
Question 3: Is it no business with the repos what I set
in /etc/yum.repos.d? I mean setting centos repo with my machine does’t
influence setup openstack? the openstack packages should come from
“OpenStack” channel, Am I right?
Question 4: how could I install the openstack packages on RH, I am
desperate for experience on it.
Best Regards,
Hongna
email message attachment
-------- Forwarded Message --------
From: Alan Pevec <apevec at redhat.com>
To: rhos-list at redhat.com
Subject: Re: [rhos-list] Some issues about install packages of Openstcak
Date: Tue, 11 Sep 2012 11:14:47 +0200
On 09/11/2012 08:48 AM, Xu, HongnaX wrote:
> And why I encounter 110 Error?
Try setting proxy_* parameters in /etc/rhsm/rhsm.conf
> is it the exactly the reason that I can’t get the packages of Openstack ?? and how to fix it?
You must first see rhel-server-ost-6-preview-rpms in "yum repolist" output.
If you still don't get it after setting proxy in rhsm.conf, please ask Red Hat customer service to verify entitlements in your RHN account.
> 6.With a second machine I subscribe “openstack” “Red Hat Beta” “Red Hat Enterprise Linux Server”, and set centos repo in /etc/yum.repos.d, and I could find the packages of dnsmasq-utils and git, Though encountered 110 Error with yum command either.
Mixing centos and RHEL is definitely not supported.
Regards,
Alan
_______________________________________________
rhos-list mailing list
rhos-list at redhat.com
https://www.redhat.com/mailman/listinfo/rhos-list
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/rhos-list/attachments/20121110/42189b84/attachment.htm>
More information about the rhos-list
mailing list