-------- Forwarding messages -------- From: "Andy Wang" <[email protected]> Date: 2016-04-20 17:44:14 To: "Rajagopalan Sivaramakrishnan" <[email protected]> Cc: "[email protected]" <[email protected]>,[email protected] Subject: Re:Re: [Users] Contrail R2.2 forward performance issue Hi Raja， If the bug reproduce again, I will use tcpdump to capture packets, I will try to reproduce it. Last time, I use ifconfig to display tap interface information, like following log: There is no drop statistics. tap85bfd0cf-7b: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet6 fe80::786f:e4ff:fe56:5c79 prefixlen 64 scopeid 0x20<link> ether 7a:6f:e4:56:5c:79 txqueuelen 500 (Ethernet) RX packets 8592742 bytes 35195359384 (32.7 GiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 1478520 bytes 57061641494 (53.1 GiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 At 2016-04-20 07:31:25, "Rajagopalan Sivaramakrishnan" <[email protected]> wrote: Hi Junwei, Can you please run tcpdump on the tap interfaces to see if there are retransmits from the sending VM? Raja From: Users <[email protected]> on behalf of Andy Wang <[email protected]> Date: Friday, April 15, 2016 at 1:22 AM To: "[email protected]" <[email protected]> Subject: [Users] Contrail R2.2 forward performance issue Hi All, Toady I met one vrotuer forwarding performance declining issue. Could some experts help me analyze it? Topology: Openstack version is kilo version , contrail version is R 2.2 and iperf version is 2.0. I uesd the 2 VMs in the same computer node to verify the vrouter forwarding performance. The VM machine is Ubuntu 12.04 OS. I initiated the iperf server in the VM A, and iperf client in the VM B. Then the TCP Traffic can get 9.0+ Gbps throughput. But sometimes TCP traffic could only 1.0+ Mbps and it could recover until I restart the supervisor-vrouter service in the computer node. I could not find the drop statics in the tap interface. I thought that the communication between the two VMs in the same computer node and there is no encapsulation and decapsulation in the forward process. If someone know the reason or could provide the thread , please let me know. Thanks Junwei Wang
_______________________________________________ Dev mailing list [email protected] http://lists.opencontrail.org/mailman/listinfo/dev_lists.opencontrail.org
