Openstack: Data network is added, floating ip is assigned but cant be reached

[BN]

Data network is added as a flat physnet0 network, screenshot: https://prnt.sc/t1gmij

Here is full info related to my current environment and openstack: http://paste.openstack.org/show/794894/

2 Controllers, OAM:
external_oam_subnet: 10.0.0.0/16
external_oam_gateway_address: 10.0.0.1
external_oam_floating_address: 10.0.10.10
external_oam_node_0_address: 10.0.10.11
external_oam_node_1_address: 10.0.10.12

In Openstack, public network is created with ip address pool 10.0.140.10,10.0.145.100. Router is created and can be pinged. Sec groups is added with icmp/tcp-all. VM's are created, floating ip assigned, machines cant be pinged/ssh'd.

Would really appreciate if you can help, please feel free to ask for any additional info... if neutron logs are required or anything else.

[Yatindra Shashi (yshashi)]

Hi,

I also had similar issue once that I was not able to ping VM.
Than I enabled in the securities group ICMP Ingress & Egress and SSH (port 22 ) separately and it worked.

Egress IPv4 ICMP Any 0.0.0.0/0
Ingress IPv4 ICMP Any 0.0.0.0/0
Ingress IPv4 TCP 22 (SSH) 0.0.0.0/0

Are you able to ping local router from VM?
Can you ping using netns:
$ sudo ip netns exec qdhcp-xxxxxxx ssh -i ~/.ssh/id_rsa ubuntu@Ip of machin

[BN (zatoichy)]

Hi,

If you open paste.openstack log its all there. Yes, I can ping router. Yes, I ve added sec groups. No, I still cant ping machines :(

[BN (zatoichy)]

Hi Ghada,

I am wondering, if I can push my mail somewhere to get answers on http://lists.starlingx.io/pipermail/starlingx-discuss/2020-June/008953.html ? Thank you

[Ghada Khalil (gkhalil)]

Hello BN,
I've asked Joseph Richard to help you trouble-shoot the issue reported here.
Regarding your email, the starlingx-discuss mailing list is the correct place to send. Hopefully, you'll get some responses shortly.
Regards,
Ghada

[Joseph Richard (josephrichard)]

Can you check inside of the VM (virsh console or nova console) if it has an IP address? Are you able to ping out from the VM?
What route are you trying to take to the VM?

You could try tcpdumping to see where the ping is getting lost,

[BN (zatoichy)]

Hi Joseph,

Okay, obviously the issue is that my cirros vm doesnt have ip address asigned to it. When I go into vm, there is no ip address. https://prnt.sc/t704ga

However, I can see its active and assigned in openstack? https://prnt.sc/t704q9

Is it possible that it can be neutron issue?

My private network is vxlan.

[Joseph Richard (josephrichard)]

I just looks at your logs again. You have enable_dhcp off on your subnet. Try turning it on on your subnet and then rebooting your guests.

[BN (zatoichy)]

Dhcp is disabled on my public network so pool is allocated and specified. In private vxlan its enabled.

[Joseph Richard (josephrichard)]

So what should be answering the dhcp request from your guest?

[BN (zatoichy)]

vxlan, private network - http://paste.openstack.org/show/795281/

[Joseph Richard (josephrichard)]

Can you try locking controller-1 to rule out a connectivity issue between the two controllers?
Can you add the output of neutron port-list?

[Ghada Khalil (gkhalil)]

This is likely a specific configuration issue. Will need more info as requested above to investigate further.

[BN (zatoichy)]

Issue can be closed. It was related to vxlan missconfiguration