Hi Raja,

I did an uninstall and setting up afresh. This time fab setup_all
installation is not successful since keystone services are not coming up.

2016-09-22 13:10:19:982340: [root@1.1.1.1] out: WARNING: Bypassing
authentication using a token & endpoint (authentication credentials are
being ignored).
2016-09-22 13:10:20:285151: [root@1.1.1.1] out: An unexpected error
prevented the server from fulfilling your request. (HTTP 500) (Request-ID:
req-0fc752e4-30a5-4b70-9601-25f727ca0b55)
2016-09-22 13:12:00:450949: [root@1.1.1.1] out: + '[' 1 == 0 ']'
2016-09-22 13:12:00:483068: [root@1.1.1.1] out: + echo 'Keystone is not
up, retrying in 5 secs'
2016-09-22 13:12:00:483259: [root@1.1.1.1] out: Keystone is not up,
retrying in 5 secs
2016-09-22 13:12:00:483359: [root@1.1.1.1] out: + sleep 5
2016-09-22 13:12:00:483446: [root@1.1.1.1] out: + for i in '{1..36}' '{'
2016-09-22 13:12:05:499173: [root@1.1.1.1] out: + keystone tenant-list
2016-09-22 13:12:05:499280: [root@1.1.1.1] out: WARNING: Bypassing
authentication using a token & endpoint (authentication credentials are
being ignored).

Does it indicate any missing config in testbed.py? Do I have to uncomment
env.keystone section? Or could it be a known issue?

The testbed.py is changed to make use of  single NIC interface for
management, control, data.

#Management ip addresses of hosts in the cluster
host1 = 'root@1.1.1.1'
host2 = 'root@1.1.1.2'

control_data = {
host1 : { 'ip': '1.1.1.1/24', 'gw' : '0.0.0.0', 'device':'eth4'},
host2 : { 'ip': '1.1.1.2/24', 'gw' : '0.0.0.0', 'device':'eth4'},
}




Thanks
Harish



>I don’t see anything unusual.
>I think before that, I need confirmation to the basic question I had.
>I’m attaching the pic again for reference.
>The host1 can reach host2 either over the 172.29.54.x network via eth0
>(DHCP’ed address) or over NIC interfaces via eth4 or eth5 (with static
>ipaddr).
>So what should ideally be the management IP address in testbed.py?
>My control_data looks like:
>control_data = {
>host1 : { ‘ip: ‘1.1.1.1/24’, ‘gw’ : 172.29.48.1’, ‘device’ : ‘eth4’ }
>host2 : { ‘ip: ‘1.1.1.2/24’, ‘gw’ : 172.29.48.1’, ‘device’ : ‘eth4’ }
>
>
>Thanks,
>harish 
>
>
>
>>Please take a look at /etc/network/interfaces to see if eth0 has the
>>right entries.
>>The Contrail installation process add the vhost0 interface to this file
>>(in place of eth0).
>>But if you hit some errors, perhaps this file was not in a sane state.
>>
>>Raja 
>>
>>On 9/21/16, 4:37 PM, "Patil, Harish" <harish.pa...@cavium.com> wrote:
>>
>>    Hi Raja,
>>    I have a strange issue.
>>    After issuing fab upgrade_kernel the kernel is downgraded to
>>    3.13.0.85-generic as expected. But after that I can’t ping/ssh the
>>eth0
>>    interface of host1. I tried all other ways to make it work. Have
>>someone
>>    else seen the same issue? Even though I boot back to 4.2.0-27-generic
>>    kernel I can’t ping. Could the contrail SW be interfering with it?
>>    
>>    Thanks,
>>    Harish
>>    
>>    
>>    >Thanks Raja.
>>    >I will try downgrading the kernel and see if I can get over this.
>>    >Can you please re-confirm that my understanding of control_data
>>section is
>>    >correct, I shall skip VLAN.
>>    >
>>    >Harish
>>    >
>>    >>Hi Harish,
>>    >>    Yes, you need “fab upgrade_kernel”. It could downgrade the
>>kernel.
>>    >>You can avoid VLAN in the control_data,
>>    >>but the gateway may be required.
>>    >>
>>    >>Raja
>>    >>
>>    >>On 9/20/16, 10:23 AM, "Users on behalf of Patil, Harish"
>>    >><users-boun...@lists.opencontrail.org on behalf of
>>    >>harish.pa...@cavium.com> wrote:
>>    >>
>>    >>    Hello,
>>    >>    I¹m trying out Contrail SW installation for the first time. I
>>have an
>>    >>    issue in installing it.
>>    >>    
>>    >>    Attached are the following to help debug the issue:
>>    >>    1) Diagram that shows the setup. Its a simple two-node topology
>>with
>>    >>a
>>    >>    2x25G NIC connected back-back.
>>    >>    The testbed.py is configured based on that. All the
>>configuration is
>>    >>    happening over the lab network 172.29.54.x connected via eth0
>>    >>interfaces.
>>    >>    
>>    >>    The end objective is to create VMs on both nodes using
>>Openstack and
>>    >>ping
>>    >>    between both VMs using VXLAN as the tunneling protocol using
>>the NIC
>>    >>    interfaces.
>>    >>    
>>    >>    2) The fab setup_all log that shows the failure.
>>    >>    
>>    >>    3) testbed.py
>>    >>    
>>    >>    Additional details:
>>    >>    
>>    >>    1) Contrail SW:
>>    >>    contrail-install-packages_3.0.2.0-51-ubuntu-14-04kilo_all.deb
>>    >>    
>>    >>    2) OS:
>>    >>    root@dut4092:/opt/contrail/utils# uname -a
>>    >>    Linux dut4092 4.2.0-27-generic #32~14.04.1-Ubuntu SMP Fri Jan
>>22
>>    >>15:32:26
>>    >>    UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
>>    >>    
>>    >>    Questions:
>>    >>    
>>    >>    - The documentation says 3.13.0-34 is the recommended kernel
>>version.
>>    >>Is
>>    >>    it a must?
>>    >>    Does fab upgrade_kernel_all actually downgrades the linux
>>kernel?  I
>>    >>    didn¹t experiment by downgrading the kernel since the current
>>    >>installation
>>    >>    issue does not seem related to linux kernel version.
>>    >>    
>>    >>    - There is a section called ³control_data² in testbed.py. I
>>believe
>>    >>this
>>    >>    is where we need to specify eth4 and eth5 as NIC interfaces
>>over
>>    >>which the
>>    >>    VM traffic has to flow.
>>    >>    But I don¹t want to specify gateway and VLAN. Is it possible?
>>    >>    Eg:
>>    >>    #    host1 : { 'ip': '192.168.10.1/24', 'gw' :
>>'192.168.10.254',
>>    >>    'device':'eth0', 'vlan': '224' },
>>    >>    
>>    >>    Thanks,
>>    >>    
>>    >>    Harish
>>    >>    
>>    >>    
>>    >>
>>    >>
>>    >
>>    >
>>    >
>>    
>>    
>>    
>>
>>
>
>
>


_______________________________________________
Users mailing list
Users@lists.opencontrail.org
http://lists.opencontrail.org/mailman/listinfo/users_lists.opencontrail.org

Reply via email to