Opened 15 years ago

Last modified 15 years ago

#5489 new defect

high-bandwidth multicast traffic

Reported by: Stefaan Owned by:
Component: network Version: VirtualBox 3.0.10
Keywords: Cc:
Guest type: Linux Host type: Linux



When I generate a lot of multicast traffic (500mbps) from some host on the network to a virtual machine with a virtual e1000 (over the software network bridge), the network on my virtual machine hangs after a couple of seconds. All hosts (virtual and non-virtual) are running Debian Lenny. The physical host gives kernel stack traces in its kernel ring buffer that seem to be related.

The physical host is a four-core Intel Core2 which can handle the multicast traffic in the host os. The virtual host is a four-core as well, with a virtual e1000.

I will attach the kernel ring buffer log.

Attachments (1)

dmesg-cut (7.5 KB ) - added by Stefaan 15 years ago.
kernel ring buffer log extract

Download all attachments as: .zip

Change History (2)

by Stefaan, 15 years ago

Attachment: dmesg-cut added

kernel ring buffer log extract

comment:1 by Stefaan, 15 years ago

Retried with Virtualbox 3.1.0.

Though I did not get the "soft lockup" messages, the network again blocks after a few seconds of high-bandwidth multicast traffic. "tcpdump" on both host on virtual machine show that traffic from the virtual machine to the host still works, but traffic from the host does not reach the virtual machine. After running "ethtool -r eth0" in the virtual machine, normal traffic seems to be restored.

Note: See TracTickets for help on using tickets.

© 2023 Oracle
ContactPrivacy policyTerms of Use