Re: [HTCondor-users] identifying (gracefully/peacefully) shutting down daemons

[Brian Bockelman <bbockelm@xxxxxxxxxxx>]

> On Aug 5, 2016, at 2:21 AM, Thomas Hartmann <thomas.hartmann@xxxxxxx> wrote:
> 
> Hi Brian,
> 
> On 2016-08-04 21:54, Brian Bockelman wrote:
>> Hi Thomas,
>> 
>> I donât know of any way to do this myself.
>> 
>> Note thereâs two levels here:
>> - Determining the state of the master.  Is it trying to shut down daemons?
> 
> yes, to read the state of the master for its daemons would be most
> interesting
> 
>> - Determining the state of the child daemons.  Did they successfully get the signal from the master?
> 
> What I am not sure about, if I got it correctly, is if it is possible to
> differentiate from which daemon another daemon got a signal. I.e., if a
> startd got into draining due to the master preparing a shutdown or due
> to defrag?

I believe, internally, draining and shutdown are slightly different states: if we had a mechanism to query the current state, then you should be able to distinguish them.

Do you have a gittrac account currently?  Can I talk you into filing a bug on the issue?

Brian