[Xen-users] All domUs locked up after massive amounts of dom0 console text

[Andy Smith <andy@xxxxxxxxxxxxxx>]

Earlier today I did something silly and added an ebtables rule to my
dom0 to log to console every single ethernet frame on any interface.
At that point my dom0 became largely unusable as a massive stream of
log messages went to its console.

Fortunately with extreme patience, I was able to painstakingly log
in over the serial console and flush the rule out.  That took about
30 minutes!

However, I then discovered that every single domU on the machine was
locked up solid.  If I connected to their consoles I would get no
input or output.  Trying to ssh to them resulted in nothing at all
happening (not even a timeout).

dom0 was behaving OK but every xm command issued returned
immediately without seeming to do anything.  I finally tried an xm
destroy and got a lot of nasty messages about eth0s still in use,
and all domains showing as using 0 RAM.

They wouldn't start again and the messages wouldn't stop coming, so
I had to reboot dom0.

I see in http://wiki.xensource.com/xenwiki/XenTodoList you have:

Bugs

 3 investigate xcs lockups under high rate console output (seperate
   console from xend)

Would that be the same issue?

signature.asc
Description: Digital signature

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users