On 02/05/13 17:32, steve_1...@hushmail.com wrote: > On Thu, May 2, 2013 at 11:17 AM, Roger Pau Monné <roger....@citrix.com> wrote: >> On 02/05/13 17:09, agya naila wrote: >>> Does any one have test this script performance with Xen live migration? >>> what is the performance result compared with brctl. I hope any body want >>> to share it. I want to use this script with my Xen 4.2.1 using xend/xm >>> toostack. >> >> I have not tried it myself, but I'm not sure this will play nice with >> xend/xm, AFAIK xend had a race regarding device destruction and hotplug >> execution, that's one of the points why we call hotplug scripts from >> libxl directly instead of calling them from udev. Any reason to use xm >> instead of xl? > > I think I should ask in separate email but anyways. Does xl with > run_hotplug_scripts=0 > (forcing hotplug through udev) has race with udev?
AFAIK the hotplug script could be called by udev when the xenstore entries are already gone, because xm and libxl (with run_hotplug_scripts=0) removes the entries from xenstore, but has no control of when udev calls the hotplug script. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev