Hello, all!
I've been playing around with netmap successfully on various local machines but am having difficulties on EC2, specifically with FreeBSD 11.
Using Collin Percival's 10.3 AMI, I'm able to add the netmap device, rebuild the kernel, and run netmap samples liked "bridge.c"
However, with the 11 AMI (which I'd like to use to make use of ixvX -- I'm using a c4.large instance), I can't even get the "bridge" sample to work. It runs but seemingly gets no traffic. In particular, my ssh connection dies. If I capture output to a file, I get the "poll timeout" messages every 2.5s until I reboot the instance.
Here's the command I'm executing before "bridge" to try to make the system netmap-friendly. It works on some of my other machines, including ones running FreeBSD 11.0-RELEASE, though they are using emX or reX interfaces:
Then I run:
Has anyone else run into this problem? Is there any other kind of logging I can do to narrow this problem down?
Thanks!
I've been playing around with netmap successfully on various local machines but am having difficulties on EC2, specifically with FreeBSD 11.
Using Collin Percival's 10.3 AMI, I'm able to add the netmap device, rebuild the kernel, and run netmap samples liked "bridge.c"
However, with the 11 AMI (which I'd like to use to make use of ixvX -- I'm using a c4.large instance), I can't even get the "bridge" sample to work. It runs but seemingly gets no traffic. In particular, my ssh connection dies. If I capture output to a file, I get the "poll timeout" messages every 2.5s until I reboot the instance.
Here's the command I'm executing before "bridge" to try to make the system netmap-friendly. It works on some of my other machines, including ones running FreeBSD 11.0-RELEASE, though they are using emX or reX interfaces:
sudo ifconfig ixv0 -rxcsum -txcsum -rxcsum6 -txcsum6 -tso -tso6 -tso4 -lro -vlanmtu -vlanhwtag -vlanhwfilter -vlanhwtso
Then I run:
bridge netmap:ixv0 netmap:ixv0
Has anyone else run into this problem? Is there any other kind of logging I can do to narrow this problem down?
Thanks!
Last edited by a moderator: