Ask Your Question

Silvia's profile - activity

2018-10-22 07:47:56 -0500 commented answer SFC not working in Ocata

Timestamps are in the wrong order. The reply do: DST-SRC-VF1-VF2. I was expecting ICMP request doing SRC - VF1- VF2- DST

2018-10-22 07:46:26 -0500 commented answer SFC not working in Ocata

There are also some strange things: - once I see the traffic in all the interfaces, if I stop the ping and restart, I see it again only at the src and the dst - In the VFs I see only reply packets (I was expecting to see only request since the setting is unidirectional from src to dst)

2018-10-22 07:46:26 -0500 received badge  Commentator
2018-10-22 07:45:27 -0500 commented answer SFC not working in Ocata

Hi, thanks for you suggestion. I moved to the Rocky version and I had the same problem. I succeeded in having traffic in the VF ports, but there are still strange things: I have to set the security group including ICMP traffic for all the ports or I have to disable the anti spoofing?

2018-10-22 03:16:47 -0500 received badge  Popular Question (source)
2018-10-18 09:34:57 -0500 asked a question Problems with SFC Devstack Ocata

Hi all, I've installed Openstack Ocata using devstack. 1 controller + compute and 2 compute nodes. Than I have followed and

to deploy a SFC. In my setup I have: 6 ports 1 src attached to port 1 1 dst attached to port 6 2 SFs respectively attached to port 2 and 3 and port 4 and 5. Everything is in the same subnet.

I have 3 problems/questions:

1: I don't know why, the VMs are instantiated only in the Controller+Compute node. Maybe a problem of Ocata?

I did a capture on the tap interface referred to the ports. I noticed that

2: Only the first packet is in the instances selected as VNF and not all the packets. At src and dst I see all the packets

3: I have checked the timestamp of the firts packet and they have a wrong order, in particular: src, dst, VNF1, VNF2

Moreover, I would like to know how to add onos as in these slides, and what Onos controlls in this case.

Is anyone able to help me?

Thanks a lot Silvia

2017-12-14 16:57:23 -0500 received badge  Popular Question (source)
2017-12-14 16:57:23 -0500 received badge  Famous Question (source)
2017-12-14 16:57:23 -0500 received badge  Notable Question (source)
2017-12-06 09:02:19 -0500 asked a question Devstack installation with Ryu

Hi all, I know that the OFAgent is deprecated but Ryu is still supported with OpenStack by using the OVSAgent. Do you have any walk-trough and/or local.conf to reproduce the setup? Thank you


2017-06-19 00:45:05 -0500 received badge  Notable Question (source)
2017-06-19 00:45:05 -0500 received badge  Famous Question (source)
2017-02-17 14:57:08 -0500 received badge  Famous Question (source)
2016-10-28 01:33:26 -0500 received badge  Popular Question (source)
2016-09-21 05:20:29 -0500 asked a question How to set SPICE with Devstack

Hi all, I'm installing Openstack using Devstack, branch Mitaka but due to this bug I'm not able to use vnc console.

I would try to use spice but I can't find reference related to how to configure it with devstack. Can you give me an example for both compute and controller node? I also tried to modify my nova.conf setting

enabled = True

and I have this error in the n-spice log when I restart the service:

2016-09-21 12:17:35.734 INFO nova.console.websocketproxy [-] WebSocket server settings:
2016-09-21 12:17:35.735 INFO nova.console.websocketproxy [-]   - Listen on
2016-09-21 12:17:35.735 INFO nova.console.websocketproxy [-]   - Flash security policy server
2016-09-21 12:17:35.735 INFO nova.console.websocketproxy [-]   - Web server (no directory listings). Web root: /usr/share/spice-html5
2016-09-21 12:17:35.735 INFO nova.console.websocketproxy [-]   - No SSL/TLS support (no cert file)
2016-09-21 12:17:35.736 CRITICAL nova [-] error: [Errno 98] Address already in use

2016-09-21 12:17:35.736 TRACE nova Traceback (most recent call last):
2016-09-21 12:17:35.736 TRACE nova   File "/usr/local/bin/nova-spicehtml5proxy", line 10, in <module>
2016-09-21 12:17:35.736 TRACE nova     sys.exit(main())
2016-09-21 12:17:35.736 TRACE nova   File "/opt/stack/nova/nova/cmd/", line 49, in main
2016-09-21 12:17:35.736 TRACE nova     port=CONF.spice.html5proxy_port)
2016-09-21 12:17:35.736 TRACE nova   File "/opt/stack/nova/nova/cmd/", line 73, in proxy
2016-09-21 12:17:35.736 TRACE nova     RequestHandlerClass=websocketproxy.NovaProxyRequestHandler
2016-09-21 12:17:35.736 TRACE nova   File "/usr/local/lib/python2.7/dist-packages/websockify/", line 973, in start_server
2016-09-21 12:17:35.736 TRACE nova     tcp_keepintvl=self.tcp_keepintvl)
2016-09-21 12:17:35.736 TRACE nova   File "/usr/local/lib/python2.7/dist-packages/websockify/", line 741, in socket
2016-09-21 12:17:35.736 TRACE nova     sock.bind(addrs[0][4])
2016-09-21 12:17:35.736 TRACE nova   File "/usr/lib/python2.7/", line 228, in meth
2016-09-21 12:17:35.736 TRACE nova     return getattr(self._sock,name)(*args)
2016-09-21 12:17:35.736 TRACE nova error: [Errno 98] Address already in use
2016-09-21 12:17:35.736 TRACE nova

It sems like it can read the nova.conf

Can anyone help me? Thanks


2016-07-20 20:06:40 -0500 received badge  Popular Question (source)
2016-07-20 20:06:40 -0500 received badge  Notable Question (source)
2016-06-07 13:48:26 -0500 received badge  Famous Question (source)
2016-05-05 04:34:41 -0500 asked a question Openvswitch agent not alive

Hi all,

I have installed OpenStack via Devstack, Liberty, I created the external network, the private one and the router to connect each other, launched instances and assigned the floating IP.

BUT when I check the connectivity I can't ping the router interface and from the instances, I can't ping an external address.

I have checked
neutron agent-list

id | agent_type | host | alive | admin_state_up | binary | +--------------------------------------+--------------------+-----------+-------+----------------+---------------------------+ | 0b1a6478-4f1d-4338-a45d-112fda0c46af | DHCP agent | devstack1 | :-) | True | neutron-dhcp-agent | | 53a726ca-d874-41da-a06e-9e30624e6fce | Open vSwitch agent | devstack1 | xxx | True | neutron-openvswitch-agent | | a8938910-8ef2-4894-a31c-5047d536156d | L3 agent | devstack1 | :-) | True | neutron-l3-agent | | ede1be74-12fc-435c-9f4b-ef44ae088a2d | Metadata agent | devstack1 | :-) | True | neutron-metadata-agent | +--------------------------------------+--------------------+-----------+-------+----------------+---------------------------+

the openswitch agent is not alive.

How to debug it?

2016-05-04 09:19:37 -0500 received badge  Notable Question (source)
2016-05-04 06:44:51 -0500 received badge  Famous Question (source)
2016-05-02 07:38:08 -0500 received badge  Popular Question (source)
2016-05-01 16:49:56 -0500 asked a question [Mitaka] Can't start instances

Hi all, I want to use openstack with Gnocchi (I need bandwidth statistics). Firstly I tried to install it through Devstack with the Liberty release, but I faced an error related to Keystone during the stacking. So I decided to move to Mitaka. The stack is completed without errors, but when I try to launch an istance (using the GUI) I can't find images and I have these errors: - Unable to retrieve the extensions. - Unable to retrieve the flavors. - Unable to retrieve the images.

I also tried to do it using the shell but, when I do "source openrc admin admin" to login there is this warning: WARNING: setting legacy OS_TENANT_NAME to support cli tools.

I am also looking for further errors in the logs, but it seems there are not other errors. But I'm going to better investigate.

Any suggestion? Is it not possible use gnocchi with liberty? I had no errors with that branch.

Thank you

2016-01-08 12:09:00 -0500 received badge  Notable Question (source)
2015-12-15 11:16:26 -0500 received badge  Popular Question (source)
2015-10-27 09:38:23 -0500 asked a question OpenDaylight and Openstack - Can't connect node to an ovs

Hi all, I'm integrating OpenStack via DevStack with OpenDaylight. I'm using DevStack Kilo and OpenDaylight Lithium. In a server I have 5 VMs:

VM1: eth0 eth1 This is my OpenStack Controller+compute.

VM2: eth0 eth1 Compute node

VM3: eth0 eth1 Compute node

VM4: eth0 ODL Controller

VM7: eth0 eth1 eth2 eth3 OpenVSwitch

I consider eth0 the mgmt network (I also use this address to access via ssh into the VMs) and eth1 the data network (VM1 eth1 is connected to OVS eth1, VM2 eth1 is connected to OVS eth2, VM3 eth1 is connected to OVS eth3). This network works properly, I've checked.

But, when I stack VM1, vm2 and VM3 I see:

  • From DLux that the virtual switches inside the compute nodes are connected each other (like a ring topology) and my OVS is isolated.

  • When I launch instances they are correctly created, but if I ping them (e.g. Instance1 in VM1 and Instance2 in VM2) the traffic goes through the mgmt network (eth0).

What I want is that the traffic passes through my data network. Can you give me some suggestions?

Thank you



2015-07-08 08:33:25 -0500 received badge  Famous Question (source)
2015-06-26 04:45:39 -0500 commented answer Error: Failed to launch instance "inst": Please try again later [Error: No valid host was found. Exceeded max scheduling attempts 3 for instance ....

I noticed that thia happen when I restart the ODL controller.

2015-06-22 02:17:39 -0500 commented answer Error: Failed to launch instance "inst": Please try again later [Error: No valid host was found. Exceeded max scheduling attempts 3 for instance ....

Hi Yaolin. Yes, nova compute is running in my compute nodes.

2015-06-22 02:11:16 -0500 received badge  Famous Question (source)
2015-06-20 10:29:43 -0500 received badge  Notable Question (source)
2015-06-19 12:27:37 -0500 received badge  Popular Question (source)
2015-06-19 07:50:22 -0500 asked a question Error: Failed to launch instance "inst": Please try again later [Error: No valid host was found. Exceeded max scheduling attempts 3 for instance ....

Hi all, I have 4 VMs in a server, managed by ESXi. One of these perform as OVS, one is the Openstack controller and the other two are the compute hosts. Yesterday I set up Openstack via devstack with Opendaylight integration for VTN. There was some issue in the VTN configuration but I was able to create the instances. This morning I tryed again to launch the instances before facing the other problems and there was this error:

Error: Failed to launch instance "inst": Please try again later [Error: No valid host was found. Exceeded max scheduling attempts 3 for instance a12ab017-396a-46d0-9749-011b829c6273. Last exception: [u'Traceback (most recent call last):\n', u' File "/opt/stack/nova/nova/compute/", line 2081, in _do_build_and_run_in].

My VMs have the maximun availability of memory and CPU (the same as yesterday). I tryed with a tiny flavour but there is the same error. When I do: nova-manage service list in my controller node I have this output:

Binary Host Zone Status State Updated_At

nova-conductor devstack1 internal enabled :-) 2015-06-19 12:45:16

nova-compute devstack1 nova enabled :-) 2015-06-19 12:45:08

nova-cert devstack1 internal enabled :-) 2015-06-19 12:45:10

nova-scheduler devstack1 internal enabled :-) 2015-06-19 12:45:10

nova-consoleauth devstack1 internal enabled :-) 2015-06-19 12:45:14

nova-compute devstack2 nova enabled :-) 2015-06-19 12:45:10

nova-compute devstack3 nova enabled :-) 2015-06-19 12:45:11

(all smiling!!)

I have also read to check the file /var/log/nova/nova.log (if I'm not wrong), but I can't find the directory nova inside /var/log

Thank you

2015-06-05 09:50:29 -0500 received badge  Famous Question (source)
2015-06-05 09:50:29 -0500 received badge  Notable Question (source)
2015-05-19 08:40:01 -0500 received badge  Popular Question (source)
2015-05-19 08:40:01 -0500 received badge  Notable Question (source)
2015-05-14 04:45:59 -0500 received badge  Editor (source)
2015-05-14 04:22:57 -0500 asked a question Openstack in a "real" setup with ODL controller

Hi all,

I would like to build a sort of real testbed to use OpenStack with an ODL Controller.

I have few machine in which I have instelled respectively:

devstack1 -> OpendayLight + Openstack controller

devstack2 and devstack3 -> Compute nodes

ovs1 -> openvSwitch to connect the controller to the compute hosts.

First of all I have downloaded ODL and installed some component. After that I was able to access to the ODL dashboard. Than I have installed the Openstack controller with Devstack and the two compute hosts.

To prepare the local.conf file I've used this guide:

I could access to the devstack dashboard BUT I can't create new instaces in the compute hosts. Moreover I can not anymore access to the ODL controller dashboard.

For some reason I had to reboot the controller machine, I have done ./rejoin-stack to access again to OpenStack, but now I can't access to the dashboard

An error occurred authenticating. Please try again later.

or, after the authentication (source openrc admin admin), if I try to do, for example, nova list it gives me this error:

Authorization Failed: An unexpected error prevented the server from fulfilling your request: (OperationalError) (2003, "Can't connect to MySQL server on '' (111)") None None (Disable debug mode to suppress these details.) (HTTP 500)

BTW... I have also installed OVSwitch because I want that my controller is connected to the compute hosts through a "real" switch OpenFlow.

I've done:

sudo ovs-vsctl set-controller openflow tcp:*controllerIP*:6633

to set the connection to the controller, how to set the connection to the hosts?

Can you help me in some way? I need that this basic configuration works ASAP.


2015-05-08 18:21:41 -0500 received badge  Popular Question (source)
2015-05-06 09:44:12 -0500 commented question Neutron-legacy error

I'm following this guide:

2015-05-06 08:52:38 -0500 asked a question Neutron-legacy error

Hi all, I'm installig Openstack using Devstack, and I'm includind the ODL plugin.

This is my local.conf file:


#IP Details
HOST_IP= #Please Add The Control Node IP Address in this line

#Instance Details
#config Details
RECLONE=yes #Make it "no" after stacking successfully the first time
#OFFLINE=True #Uncomment this after stacking successfully the first time


#ML2 Details
disable_service n-net
enable_service q-svc
enable_service q-meta
enable_service q-dhcp
enable_service q-l3
enable_service neutron

enable_service odl-compute

ODL_MGR_IP= #Please Add the ODL IP Address in this line

#Details of the Control node for various services
url= #Please Add the ODL IP Address in this line

#set for live migration

I run ./stack and I receive this error:

2015-05-06 13:19:23.940 | + echo  'screen -t q-svc bash'  
2015-05-06 13:19:23.940 | + echo 'stuff "python /usr/local/bin/neutron-server --config-file /etc/neutron/neutron.conf --config-file /etc/neutron"'lugins/ml2/ml2_conf.ini
2015-05-06 13:19:23.941 | + [[ -n /opt/stack/logs ]]  
2015-05-06 13:19:23.941 | + echo 'logfile /opt/stack/logs/q-svc.log.2015-05-06-150747'
2015-05-06 13:19:23.941 | + echo 'log on'  
2015-05-06 13:19:23.941 | + screen  -S stack -p q-svc -X stuff 'python /usr/local/bin/neutron-server --config-file /etc/neutron/neutron.conf  --config-file /etc/neutron/plugins/ml2/ml2_conf.ini  & echo $! /opt/stack/status/stack/; fg || echo "q-svc failed to start" | tee "/opt/sta'k/status/stack/q-svc.failure"
2015-05-06 13:19:23.949 | + echo 'Waiting for Neutron to start...'
2015-05-06 13:19:23.949 | Waiting for Neutron to start...  
2015-05-06 13:19:23.949 | +is_ssl_enabled_service neutron
2015-05-06 13:19:23.950 | + local services=neutron  
2015-05-06 13:19:23.950 | + local service=
2015-05-06 13:19:23.950 | + '[' False == False ']' 
2015-05-06 13:19:23.951 | + return 1 
2015-05-06 13:19:23.951 | + timeout 60 sh -c 'while ! wget  --no-proxy -q -O-; do sleep 1; done'
2015-05-06 13:20:23.957 | + die 694 'Neutron did not start'  
2015-05-06 13:20:23.957 | + local exitcode=0
2015-05-06 13:20:23.958 | [Call Trace]
2015-05-06 13:20:23.958 | ./
2015-05-06 13:20:23.958 | /home/silvia/devstack/lib/neutron-legacy:694:die
2015-05-06 13:20:24.298 | [ERROR] /home/silvia/devstack/lib/neutron-legacy:694 Neutron did not start  
2015-05-06 13:20:25.306 | Error ...
2015-04-29 02:23:32 -0500 received badge  Famous Question (source)
2015-04-23 02:18:52 -0500 received badge  Enthusiast
2015-04-22 06:05:30 -0500 received badge  Notable Question (source)
2015-04-22 05:52:53 -0500 commented question Cinder type-create error during Devstack installation

I have also tried to install it ad a controller node in one of the machine that was a compute host, previously changing the local.conf file, but I had the same error. This could be caused by the NFS shared volume? If it's so, why I could install it the first time with no errors?

2015-04-22 05:21:50 -0500 commented question Cinder type-create error during Devstack installation

Is not possible recover these errors? I have installed openstack in a VM located into an external server managed by the university, I just asked for these VMs, so I can't personally reinstall the OS.