On Fri, Jun 21, 2013 at 3:12 PM, Matthew Miller mattdm@fedoraproject.org wrote:
On Fri, Jun 21, 2013 at 08:55:50AM +0200, Juerg Haefliger wrote:
Works for me in OpenStack. Good job Matt! Just one question, can you explain the rational behind the console crazyness?
serial=tty0 console=ttyS0,115200n8 console=hvc0 console=tty1
For OpenStack, I'd like to see something like the following so that I get console messages on the serial port which can be retrieved via a Nova call. Helps with identifying boot issues in the absence of a 'real' virtual console.
console=tty1 console=ttyS0
Does the crazy line above _not_ work for you in OpenStack? It does for me.
Only partially. In only get the kernel messages in the console log but none of the other console messages, i.e., init script outputs.
...Juerg
I posted about this a few days ago, but to recap, the simple line causes cloud-init (which wants to write directly to the console) to freak out in EC2, where ttyS0 isn't a device. Putting the hvc0 console line first causes it to _not_ freak out, and in my quick testing console output still worked in OpenStack too.
The alternative is to write different things to menu.lst for pvgrub and extlinux.conf for KVM (or other HVM) use. But that seemed a little too "magical".
-- Matthew Miller ☁☁☁ Fedora Cloud Architect ☁☁☁ mattdm@fedoraproject.org _______________________________________________ cloud mailing list cloud@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/cloud