Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-26 Thread Murali R
Correction: not Sep 22, its Oct 22. On Tue, Oct 27, 2015 at 6:15 AM, Murali R wrote: > > ​Turns out the only difference between the timing of controller node and > compute node installs was not using main ovs brach until contrack is > merged. Looks like when I installed compute (remote node) af

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-26 Thread Murali R
​Turns out the only difference between the timing of controller node and compute node installs was not using main ovs brach until contrack is merged. Looks like when I installed compute (remote node) after Sep 22, I had the right repo and it installed vport-geneve. When I build+install the old chec

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-25 Thread Murali R
Sorry for the delay, I was on flight writing my previous response and had difficulty reaching my setup in rax from Shanghai, Re-attached the sb-entries file that earlier got lost when mail group was dropped. That dump was taken from primary controller node, so not sure how the tunnels look defined

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-23 Thread Han Zhou
Hi On Fri, Oct 23, 2015 at 1:20 PM, Gurucharan Shetty wrote: > I did not read the entire thread, so here is an advice if not already > offered . If the tunnels are "geneve", make sure that 'lsmod | grep > geneve" has an entry. > The datapath port is created, which indicates kernel module should

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-23 Thread Gurucharan Shetty
I did not read the entire thread, so here is an advice if not already offered . If the tunnels are "geneve", make sure that 'lsmod | grep geneve" has an entry. On Fri, Oct 23, 2015 at 1:14 PM, Murali R wrote: > Thanks Han. I will check again, but last time no packets were going out of > the switc

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-23 Thread Murali R
Thanks Han. I will check again, but last time no packets were going out of the switch port. I think something in network config is missing, not sure what. The tunnel itself is created alright but for ovn there is an extra element of vlan mapping for network isolation through the tinnel, but don't k

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-23 Thread Han Zhou
Sorry that I dropped the list. It looks the tunnel is good. I would recommend to look into L2 before mixing with qrouter: 1. tcpdump on each tap when ping between the 2 VMs on different nodes, find out where the packet (ARP or ICMP) is dropped. 2. ovs-dpctl dump-flow on the node that packet is dro

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-23 Thread Murali R
Han, thanks for suggestions. Yes I had tried all that. Controller means node running northd. The issue is databases are all looking ok and so are tunnels they have tunnel address correct on both sides. But the overlay network created is not reachable in remote node. Host routing is not applicable h

Re: [ovs-discuss] [ovn] multi node ovn debugging

2015-10-22 Thread Han Zhou
Hi Murali, On Thu, Oct 22, 2015 at 4:42 PM, Murali R wrote: > I am trying to debug the multi node ovn setup. I see the tunnels are > created, the associations seems to be done (port, datapath etc) in the db - > however the overlay network (10.0.0.0/24) is accessible only from the node > running

[ovs-discuss] [ovn] multi node ovn debugging

2015-10-22 Thread Murali R
I am trying to debug the multi node ovn setup. I see the tunnels are created, the associations seems to be done (port, datapath etc) in the db - however the overlay network (10.0.0.0/24) is accessible only from the node running ovn-controller (even after attaching floating ip through neutron). Ev