Ask Your Question

Steven Su's profile - activity

2019-05-10 09:49:48 -0600 received badge  Famous Question (source)
2016-02-23 21:51:28 -0600 received badge  Famous Question (source)
2016-02-19 05:11:45 -0600 received badge  Notable Question (source)
2016-01-31 19:28:02 -0600 answered a question Unable to SSH to Devstack Instance from an external computer

SSH and Telnet are using different port. Telnet works doesn't mean SSH is fine. Some questions may help to investigate: - Have you enabled SSH in your project security groups? - Can yo SSH from you local host to VM? - Cirros image should allow host authentication - can you try it instead of key file?

2015-12-11 06:42:16 -0600 received badge  Notable Question (source)
2015-11-19 12:23:14 -0600 received badge  Popular Question (source)
2015-11-18 21:40:26 -0600 commented answer How does neutron node update its agents to neutron DB?

The Neutron controller node has DB connection (see http://docs.openstack.org/kilo/instal...), the neutron agent nodes don't - as they do not need to connect to DB directly.

2015-11-18 21:32:16 -0600 answered a question Connection lost between OVS patch ports?

Ok I rebuild the environment and there is no such problem, it's a wired issue but I guess I will let it go for now.

2015-11-18 21:29:58 -0600 answered a question How does OVS GRE tunnel port get network package?

Ok I get it, it's the iptables blocked the ICMP which cause the package not reaches to GRE tunnel port. In my iptables there are rules:

-A INPUT -j REJECT --reject-with icmp-host-prohibited

-A FORWARD -j REJECT --reject-with icmp-host-prohibited

Remove those rules then it works fine.

2015-11-12 11:00:50 -0600 received badge  Student (source)
2015-11-07 02:16:16 -0600 received badge  Popular Question (source)
2015-11-05 06:12:33 -0600 commented answer How does OVS GRE tunnel port get network package?

Thanks, that's one of doc in my bookmark:) But it didn't explain how package transfer from NIC to GRE tunnel port and who's control it. In my env I saw packages on eth1 but nothing on GRE tunnel port. Use "watch ovs-ofctl dump-flows br-tun" the incoming package number is zero.

2015-11-05 05:53:18 -0600 commented answer How does neutron node update its agents to neutron DB?

Yes, if you look at the neutron.conf on controller node you will find it has same Message Queue server defined. Note that file has DB connection as well so it able to update Database.

2015-11-05 00:57:46 -0600 answered a question Cinder with IBM DS6000 via Fibre Channel

I guess it lists DS8000 because that is only one tested.

You'd better contact IBM support for clarification.

Good luck!

2015-11-05 00:54:08 -0600 answered a question i am not able to log in with my host ip given after the installation of openstack,keystone is working good,i am a beginner,please help!!

You'd better give more details of your problem or otherwise no one should understand what happened.

what is your deployment structure?

what host you referred to?

what error message you have?

etc..

Good luck!

2015-11-05 00:47:26 -0600 asked a question How does OVS GRE tunnel port get network package?

I'm debugging a network issue and like to know the magic behind OVS GRE tunnel.

Let's say a GRE tunnel between host A and B, on host A I have:

Bridge br-tun
        Port "gre-0a0a0114"
            Interface "gre-0a0a0114"
                type: gre
                options: {df_default="true", in_key=flow, local_ip="10.10.1.21", out_key=flow, remote_ip="10.10.1.20"}

Host A has NIC eth1 using IP 10.10.1.21.

and on Host B I have:

Bridge br-tun
Port "gre-0a0a0115"
    Interface "gre-0a0a0115"
        type: gre
        options: {df_default="true", in_key=flow, local_ip="10.10.1.20", out_key=flow, remote_ip="10.10.1.21"}

Host B has NIC eth1 using IP 10.10.1.20.

When a package sent out from host A to B, will it first reach eth1 on Host B? Or will it reach Port "gre-0a0a0115" on br-tun directly? If the package reaches eth1 first then how does port "gre-0a0a0115" get the traffic?

I'm asking this because I found somehow the traffic reaches eth1 on host B but nothing on GRE tunnel port "gre-0a0a0115" in my environment.

Thanks in advance!

2015-11-05 00:47:26 -0600 answered a question How does neutron node update its agents to neutron DB?

I think it's using message queue. E.g. In neutron.conf it has:

rpc_backend = rabbit

If you want dig it more you may have a look at: https://www.openstack.org/assets/presentation-media/insideneutron2.pdf (https://www.openstack.org/assets/pres...)

2015-11-05 00:47:24 -0600 asked a question Connection lost between OVS patch ports?

I'm setting up Kilo KVM region and got a problem on network configuration. I have been stuck in there for a couple of days now, so any suggestions are much appreciated!

This is three a nodes environment:

kilocontroller (management network IP 192.168.155.188 on eth0)

kilonetwork (management and external network IP 192.168.155.189 on br-ex, Data network IP 10.10.1.20 on eth1)

kilocompute (management network IP 192.168.155.190 on eth0, Data network IP 10.10.1.21 on eth1)

-All notes are VMs running RHEL v7.1 on vCenter v5.5.

-I’m following the KILO RHEL installation guide (http://docs.openstack.org/kilo/install-guide/install/yum/content/ (http://docs.openstack.org/kilo/instal...)). So both tenant and tunnel use GRE.

I'm able to provision VM successfully but VM can't get IP address, by tracking the DHCP request I found the communication lost between OVS patch ports in compute node.

For example in case of VM VM-20151104-2 , the DHCP communication chain on compute node should be:

VM-20151104-2 -> tapb2517833-13 (qbrb2517833-13) qvbb2517833-13 -> qvob2517833-13 (br-int) patch-tun -> patch-int (br-tun) gre-0a0a0114 -> eth1 ====> Network Node ......

While I saw the DHCP request on port tapb2517833-13, qvbb2517833-13, qvob2517833-13, patch-tun but it didn't reach patch-int. At the same time I saw request on eth1 on both kilocompute and kilonetwork nodes (it never goes further on network node).

So the questions are:

1) Why DHCP lost between OVS patch ports? (Yes OVS peer is there)

2) How come eth1 on kilocompute get the traffic if OVS patch connection is broken?

Here is what I have on kilocompute:

[root@kilocompute ~]# brctl show
bridge name bridge id       STP enabled interfaces
qbrb2517833-13      8000.ce00be78fda3   no      qvbb2517833-13
                            tapb2517833-13

[root@kilocompute ~]# ovs-vsctl show
ff6f0cf5-f7a9-4572-89cd-4969f66e7516
    Bridge br-tun
        fail_mode: secure
        Port br-tun
            Interface br-tun
                type: internal
        Port "gre-0a0a0114"
            Interface "gre-0a0a0114"
                type: gre
                options: {df_default="true", in_key=flow, local_ip="10.10.1.21", out_key=flow, remote_ip="10.10.1.20"}
        Port patch-int
            Interface patch-int
                type: patch
                options: {peer=patch-tun}
    Bridge br-int
        fail_mode: secure
        Port patch-tun
            Interface patch-tun
                type: patch
                options: {peer=patch-int}
        Port "qvob2517833-13"
            tag: 1
            Interface "qvob2517833-13"
        Port br-int
            Interface br-int
                type: internal
    ovs_version: "2.3.1"
[root@kilocompute ~]# 


[root@kilocompute ~]# ovs-ofctl dump-ports br-int
OFPST_PORT reply (xid=0x2): 3 ports
  port LOCAL: rx pkts=0, bytes=0, drop=0, errs=0, frame=0, over=0, crc=0
           tx pkts=20, bytes=4656, drop=0, errs=0, coll=0
  port  2: rx pkts=0, bytes=0, drop=0, errs=0, frame=0, over=0, crc=0
           tx pkts=20, bytes=4656, drop=0, errs=0, coll=0
  port  3: rx pkts=20, bytes=4656, drop=0, errs=0, frame=0, over=0, crc=0
           tx pkts=0, bytes=0, drop=0, errs=0, coll=0
[root@kilocompute ~]#


[root@kilocompute ~]# ovs-ofctl dump-flows br-int
NXST_FLOW reply (xid=0x4):
 cookie=0x0, duration=40163.854s, table=0, n_packets=20, n_bytes=4656, idle_age=37601, priority=1 actions=NORMAL
 cookie=0x0, duration=40163.770s, table=23, n_packets=0, n_bytes=0, idle_age=40163 ...
(more)