ovs-vswitchd log: connection attempt failed (Address family not supported by protocol)

[Thiago Martins]

Guys,

 I'm seeing an error right after first steps towards enabling OpenvSwitch with DPDK on Xenial, easy to reproduce:

--
 NOTE: Instructions presented on an Intel post, also face this problem:

 https://software.intel.com/en-us/articles/using-open-vswitch-with-dpdk-on-ubuntu
--

 Steps to reproduce:

 * Installing

 apt update
 apt install dpdk openvswitch-switch-dpdk

 update-alternatives --set ovs-vswitchd /usr/lib/openvswitch-switch-dpdk/ovs-vswitchd-dpdk

 * Configuring

 - Set DPDK_OPTS in "/etc/default/openvswitch-switch" file:

DPDK_OPTS='--dpdk -c 0x3 -n 4'

 - Set NR_2M_PAGES in "/etc/dpdk/dpdk.conf"

NR_2M_PAGES=64

 Then, just reboot the box, to make sure that it will come enabled.

 You can see that "ps axf | grep dpdk" will show that openvswitch is, apparently, working with DPDK! But, is it really working?

 The BUG, run "ovs-vswitchd log", look:

---
root@xenial-1:~# ovs-vswitchd log
2016-03-08T05:08:35Z|00001|ovs_numa|INFO|Discovered 4 CPU cores on NUMA node 0
2016-03-08T05:08:35Z|00002|ovs_numa|INFO|Discovered 1 NUMA nodes and 4 CPU cores
2016-03-08T05:08:35Z|00003|reconnect|INFO|log: connecting...
2016-03-08T05:08:35Z|00004|reconnect|INFO|log: connection attempt failed (Address family not supported by protocol)
2016-03-08T05:08:35Z|00005|reconnect|INFO|log: waiting 1 seconds before reconnect
2016-03-08T05:08:36Z|00006|reconnect|INFO|log: connecting...
2016-03-08T05:08:36Z|00007|reconnect|INFO|log: connection attempt failed (Address family not supported by protocol)
2016-03-08T05:08:36Z|00008|reconnect|INFO|log: waiting 2 seconds before reconnect
^C2016-03-08T05:08:36Z|00009|fatal_signal|WARN|terminating with signal 2 (Interrupt)
---

Someone else is also facing this problem:

https://software.intel.com/en-us/forums/networking/topic/611366

And I strong believe that because of this problem, it is not possible to use a DPDK 10G NIC device, with OpenvSwitch+DPDK, thread on Ubuntu Devel mail list (I'm guessing here):

https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2016-March/016287.html

I really appreciate any help!

Thanks!
Thiago

[Thiago Martins (martinx)]

This is weird...

I just removed OVS with DPDK:

apt purge libdpdk0:amd64 dpdk openvswitch-switch-dpdk
apt autoremove

And installed regular OVS:

apt install openvswitch-switch

But, same problem on on log:

---
root@blade:~# ovs-vswitchd log
2016-03-08T05:57:49Z|00001|ovs_numa|INFO|Discovered 4 CPU cores on NUMA node 0
2016-03-08T05:57:49Z|00002|ovs_numa|INFO|Discovered 1 NUMA nodes and 4 CPU cores
2016-03-08T05:57:49Z|00003|reconnect|INFO|log: connecting...
2016-03-08T05:57:49Z|00004|reconnect|INFO|log: connection attempt failed (Address family not supported by protocol)
2016-03-08T05:57:49Z|00005|reconnect|INFO|log: waiting 1 seconds before reconnect
2016-03-08T05:57:50Z|00006|reconnect|INFO|log: connecting...
2016-03-08T05:57:50Z|00007|reconnect|INFO|log: connection attempt failed (Address family not supported by protocol)
2016-03-08T05:57:50Z|00008|reconnect|INFO|log: waiting 2 seconds before reconnect
2016-03-08T05:57:52Z|00009|reconnect|INFO|log: connecting...
2016-03-08T05:57:52Z|00010|reconnect|INFO|log: connection attempt failed (Address family not supported by protocol)
2016-03-08T05:57:52Z|00011|reconnect|INFO|log: waiting 4 seconds before reconnect
^C2016-03-08T05:57:52Z|00012|fatal_signal|WARN|terminating with signal 2 (Interrupt)
---

So, this might not be even a problem after all?

What is this message on log?

Thanks!
Thiago

[Thiago Martins (martinx)]

You know what, this is not even a bug! Sorry about the buzz...

"ovs-vswitch log" is wrong, "log" is not even an option of "ovs-vswitchd"!

Look:

---
root@xenial-1:~# ovs-vswitchd blsdadsa
2016-03-08T06:44:10Z|00001|ovs_numa|INFO|Discovered 4 CPU cores on NUMA node 0
2016-03-08T06:44:10Z|00002|ovs_numa|INFO|Discovered 1 NUMA nodes and 4 CPU cores
2016-03-08T06:44:10Z|00003|reconnect|INFO|blsdadsa: connecting...
2016-03-08T06:44:10Z|00004|reconnect|INFO|blsdadsa: connection attempt failed (Address family not supported by protocol)
2016-03-08T06:44:10Z|00005|reconnect|INFO|blsdadsa: waiting 1 seconds before reconnect
2016-03-08T06:44:11Z|00006|reconnect|INFO|blsdadsa: connecting...
2016-03-08T06:44:11Z|00007|reconnect|INFO|blsdadsa: connection attempt failed (Address family not supported by protocol)
2016-03-08T06:44:11Z|00008|reconnect|INFO|blsdadsa: waiting 2 seconds before reconnect
^C2016-03-08T06:44:11Z|00009|fatal_signal|WARN|terminating with signal 2 (Interrupt)
---

Also, my log file have nothing wrong:

/var/log/openvswitch/ovs-vswitchd.log:

---
2016-03-08T06:43:15.842Z|00002|vlog|INFO|opened log file /var/log/openvswitch/ovs-vswitchd.log
2016-03-08T06:43:15.892Z|00003|ovs_numa|INFO|Discovered 4 CPU cores on NUMA node 0
2016-03-08T06:43:15.892Z|00004|ovs_numa|INFO|Discovered 1 NUMA nodes and 4 CPU cores
2016-03-08T06:43:15.892Z|00005|reconnect|INFO|unix:/var/run/openvswitch/db.sock: connecting...
2016-03-08T06:43:15.893Z|00006|reconnect|INFO|unix:/var/run/openvswitch/db.sock: connected
2016-03-08T06:43:15.895Z|00007|netdev_linux|INFO|tun0 device has unknown hardware address family 65534
2016-03-08T06:43:15.896Z|00008|bridge|INFO|ovs-vswitchd (Open vSwitch) 2.5.0
2016-03-08T06:43:31.333Z|00009|memory|INFO|37520 kB peak resident set size after 16.1 seconds
---

So, I am closing this "bug" as invalid.

-
Thiago