Mailing List Archives
Authenticated access
|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] Re: NEVER MIND Condor Bug: Condor Status reatins old machine info after changing OS type.
- Date: Thu, 07 Apr 2005 14:31:47 +0100
- From: John Horne <John.Horne@xxxxxxxxxxxxxx>
- Subject: Re: [Condor-users] Re: NEVER MIND Condor Bug: Condor Status reatins old machine info after changing OS type.
On Thu, 2005-04-07 at 02:50 -0500, Derek Wright wrote:
> On Tue, 05 Apr 2005 15:11:39 +0100 John Horne wrote:
>
> > So for an 'execute-only' workstation, just running condor_startd and
> > condor_master, would sending a kill -TERM to condor_master be okay, or
> > should I send it to both the startd and the master? Using 'condor_off'
> > is not possible because the client is logged in as an unpriviledged (?)
> > user.
>
> SIGTERM to the master is totally equivalent to "condor_off -master".
> in both cases, the master will shutdown all of its children, and then
> exit. SIGTERM to both the master and the startd is unnecessary, and
> introduces a potential race condition... if the startd manages to exit
> before the master processes the SIGTERM, the master would restart the
> startd (and then kill it again, once it got the SIGTERM). so, you
> might as well just send SIGTERM to the master.
>
Many thanks, that answers my question perfectly.
John.
--
---------------------------------------------------------------
John Horne, University of Plymouth, UK Tel: +44 (0)1752 233914
E-mail: John.Horne@xxxxxxxxxxxxxx Fax: +44 (0)1752 233839