On Mon, Nov 24, 2008 at 11:53 AM, Daniel Gruner <[EMAIL PROTECTED]> wrote:
>
> Hi All,
>
> It was fun meeting you all in Austin...
>
> Now to my question:  Is there a clean way to reboot the compute nodes?
>  Here is what I have been doing:
>
> xrx -l n0000 /sbin/reboot
>
> While this works, it causes xrx to segfault:
>
> Nov 24 13:46:23 shemesh kernel: xrx[15549]: segfault at 2b0 ip 40f416
> sp 7fff1de02e00 error 4 in xrx[400000+1c000]
>
> Also, after xrx segfaults, the next character typed at the shell is
> "eaten", so the shell appears not to respond.  For example, it takes 2
> <cr> in order to get one.  This could be an issue in scripts that need
> to reboot nodes, although I haven't tested this yet.
>

sounds like on xrx process is not handling the remote xrx dying. libspfs bug?

And the 'read from keyboard' guy is still there.

wow, ow.

ron

Reply via email to