Re: capture VM network traffic

2024-12-20 Thread Jakob Bohm via 

Hi,

If the network settings on the qemu command line sends the traffic 
through a Linux kernel "bridge" device such as br0, running the capture 
tool on the host watching that bridge device should capture everything.


Each of the other network attachment options will need a different 
configuration, but in general, running the capture tool outside the VM 
prevents programs inside the VM from interfering .


If you want the capture tool to see only the VM traffic, setting up a 
dedicated bridge device for just that VM may help, but something else on 
the HM needs to carry the traffic into the world, perhaps setting up the 
HM as a router between the single VM network and the real network .


On 12/20/2024 10:46:34, lacsaP Patatetom wrote:

hi,

what network configuration (HM/VM side) would you set up to 
capture/follow a virtual machine's network traffic, forcing it to pass 
through tools such as tshark/wireshark, dnsmasq or mitmproxy ?


the host machine (HM) is an ArchLinux.

regards.


Enjoy

Jakob
--
Jakob Bohm, CIO, Partner, WiseMo A/S.  https://www.wisemo.com
Transformervej 29, 2860 Søborg, Denmark.  Direct +45 31 13 16 10
This public discussion message is non-binding and may contain errors.
WiseMo - Remote Service Management for PCs, Phones and Embedded

capture VM network traffic

2024-12-20 Thread lacsaP Patatetom 
hi,

what network configuration (HM/VM side) would you set up to capture/follow
a virtual machine's network traffic, forcing it to pass through tools such
as tshark/wireshark, dnsmasq or mitmproxy ?

the host machine (HM) is an ArchLinux.

regards.

RBD timeout 0 and QEMU process stuck during migration

2024-12-20 Thread Jiatong Shen 
Hello Experts,

   We are using QEMU 6.2 and Ceph Pacific on production. The ceph mon
and osd operation timeout is defaulting to 0 which IMO means never
timeout. Under this setup, we observe that during live migration if
the RBD network is unstable, the virtual machine could become
unresponsive (for example, virsh domjobinfo command is stuck
indefinitely).
   My question is is there any recommendation for this situation? can
I possibly cancel live migration under such circumstances?


-- 

Best Regards,

Jiatong Shen