New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
QEMU vm on xcon hub stops receiving packets #238
Comments
@plajjan as discussed, a dedicated issue for this behavior, if need be I can share the entire setup with you, but I assume this should be reproducible with any 3 VMs and sufficient application traffic |
thanks for looking into this a bit more. today i was playing with the patch in #188 and made it work with vr-bgp. i tested it all manually and didnt yet try running tests/sending a bunch of packets, but it "works"....i still need to write vr-bgp code for using native docker networking, but, i suspect if we can implement this it will work much better than the xcon/hub implementation. |
slight follow up...i tested a dot1q interface to 2 vr-bgp instances on 2 different vlans, as well as a "LAN"/"IX" interface with 2 vr-bgp instances on the lan.
|
I think there are multiple courses of action here;
|
i have been playing with the native docker networking patch. i wonder if we should just scrap xcon and use that stuff. i have it working with vr-bgp and topomachine as of earlier today. there are some missing bits to that patch, but, i am working on doing a complete implementation out of neccessity(sros mainly). |
A related issue to #231
I'm building a
vr-xcon
TcpHub
topology with 3 VMs part of the Cisco SD-WAN solution on the bridge:vmanage
vsmart
vbond
I bring up the topology and everything works fine. I can ping between the 3 VMs without issue. I now start the registration process between the
vmanage
and thevsmart
. The proprietary protocol exchange results in a stream of packets between those two components. More often than not, the outcome is that thevbond
is no longer ping-able. Important to note here is that because this is a hub, thevbond
is also receiving those packets from the hub.I always see that the
vbond
QEMU instance closes the TCP connection towards the hub. I first fixed the hub to re-build any failed connections in #237. This means that the hub will rather quickly re-establish a socket with the failed QEMU. But I'm observing that despite the TCP connection being re-established, thevbond
does not receive any of my ping requests inside the VM despite them arriving inside the container.tcpdump
output taken from thevbond
taken around the time of the outage (i.e. time where thevmanage
andvsmart
started communicating:Communication can be restored by either:
This is reproducible with QEMU 4.2.1 running all 3 containers.
The text was updated successfully, but these errors were encountered: