On Sun, Mar 31, 2013 at 6:42 PM, Michael McCandless < [email protected]> wrote:
> On Sun, Mar 31, 2013 at 6:23 PM, Robert Muir <[email protected]> wrote: > > hmm there goes that theory. Maybe its just a leftover process that didn't > > get killed from a previous smoketester: I think to be safe the python > code > > should always terminate the server it starts with 'kill -9' and nothing > > else! > > It does kill the server ... but there was a bug in that logic such > that if the 30 minute startup wait elapsed it failed to kill it. > > Which is curious ... it seems to mean that the server took 30 minutes, > didn't seem to start, but did in fact start (after 30 minutes) and > bind to the port. > maybe its related to the blackhole :)
