scp stalled on 2112KB when using internal networks with OpenVSwitch (ovs)

asked 2018-01-24 05:40:21 -0500

michelgokan gravatar image

I have 4 nodes, 1 controller (which is also act as network node), 1 cinder, 2 compute nodes. I'm using OVS and latest version of OpenStack Pike.

When I create a provider network using following command, and connect 2 vms and try to transfer a large file (using scp) everything works fine and okay:

$ ovs-vsctl add-br br-provider
$ ovs-vsctl add-port br-provider ens192
$ openstack network create --share --provider-physical-network provider --provider-network-type flat provider1
$ openstack subnet create --subnet-range 172.16.16.0/20 --gateway 172.16.16.1 --network provider1 --allocation-pool start=172.16.16.230,end=172.16.16.250 --dns-nameserver 8.8.4.4 provider1-v4

$ openstack subnet create --subnet-range fd00:172:16:16::/64 --gateway fd00:172:16:16::1 --ip-version 6 --ipv6-address-mode slaac --network provider1 --dns-nameserver 2001:4860:4860::8844 provider1-v6

$ openstack network set --external provider1

But when I create a self-service network using following commands, then I can't transger a large file between 2 vms that I connect to this self-service1 network. I can ssh and run commands with no problem but when I try to send large file (more than 2112KB) using scp command, it sends exactly 2112KB and then freeze and go into -stalled- mode:

$ openstack network create selfservice1
$ openstack subnet create --subnet-range 192.168.1.0/24 --network selfservice1 --dns-nameserver 8.8.4.4 selfservice1-v4
$ openstack subnet create --subnet-range fd00:192:168:1::/64 --ip-version 6 --ipv6-ra-mode slaac --ipv6-address-mode slaac --network selfservice1 --dns-nameserver 2001:4860:4860::8844 selfservice1-v6
$ openstack router create router1
$ openstack router add subnet router1 selfservice1-v4
$ openstack router add subnet router1 selfservice1-v6
$ neutron router-gateway-set router1 provider1

I also tried it without using a router and I have the same issue with also a simple network, when creating with a simple command openstack network create blah_blah.

I did a search over the internet and it seems this 2112KB thing is related to MTU but I set all MTUs to 1500 but still have this problem.

For more info:

ifconfig in my controller node:

ens160    Link encap:Ethernet  HWaddr 00:50:56:9a:de:70  
          inet addr:172.16.16.110  Bcast:172.16.31.255  Mask:255.255.240.0
          inet6 addr: fe80::250:56ff:fe9a:de70/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:1381623 errors:0 dropped:9 overruns:0 frame:0
          TX packets:823632 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:553791086 (553.7 MB)  TX bytes:2012204289 (2.0 GB)

ens192    Link encap:Ethernet  HWaddr 00:50:56:9a:52:93  
          inet6 addr: fe80::250:56ff:fe9a:5293/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:4296503 errors:0 dropped:0 overruns:0 frame:0
          TX packets:12115 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:2729769204 (2.7 GB)  TX bytes:880226 (880.2 KB)

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope ...
(more)
edit retag flag offensive close merge delete