Re: [condor-users] invalid sessions causing job preemption

[Todd Tannenbaum <tannenba@xxxxxxxxxxx>]

At 04:13 PM 2/9/2004, Scott Koranda wrote:
> > At 03:42 PM 2/9/2004, Scott Koranda wrote:
> > >Hi,
> > >
> > >On a Condor 6.6.0 pool running on top of RedHat 9 we are seeing
> > >the following in some of the StartLogs:
> >
> >
> > For starters:
> > Are you positive that the schedd that had the startd claimed was also
> > v6.6.0?  Or is it possible that the schedd was an older version of Condor?
>
> Yes, it does look like the schedd was from a 6.5.x. So I assume that
>
> http://www.cs.wisc.edu/~lists/archive/condor-users/msg00534.html
>
> applies.

Sure.  Or even more to the root of the trouble, I'd say that
  http://www.cs.wisc.edu/~lists/archive/condor-users/msg00553.html
applies.  If your installation does not require Condor to perform secure 
communication channels (i.e. IP based authorization is good enough for your 
setup), then I'd recommend setting
    SEC_DEFAULT_NEGOTIATION = NEVER
in all of your condor_config files and thereby bypass any/all issues 
dealing with communication session keys.


> Sorry for the trouble...

No trouble at all.

best regards Scott,
Todd


> Scott
>
> >
> > Thanks,
> > Todd
> >
> >
> > >2/9 10:33:37 DC_AUTHENTICATE: attempt to open invalid session
> > >node34:6631:1076347765:1271, failing.
> > >2/9 10:38:38 DC_AUTHENTICATE: attempt to open invalid session
> > >node34:6631:1076347765:1271, failing.
> > >2/9 10:43:37 State change: claim timed out (condor_schedd gone?)
> > >2/9 10:43:37 Changing state and activity: Claimed/Busy ->
> > >Preempting/Killing
> > >2/9 10:43:38 Got ALIVE while in Preempting state, ignoring.
> > >2/9 10:43:49 DC_AUTHENTICATE: attempt to open invalid session
> > >node34:6631:1076347835:1272, failing.
> > >2/9 10:43:49 Starter pid 32656 exited with status 0
> > >2/9 10:43:49 State change: starter exited
> > >2/9 10:43:49 State change: No preempting claim, returning to owner
> > >2/9 10:43:49 Changing state and activity: Preempting/Killing -> Owner/Idle
> > >2/9 10:43:49 State change: IS_OWNER is false
> > >2/9 10:43:49 Changing state: Owner -> Unclaimed
> > >
> > >So it looks like some problem with an invalid session is causing jobs
> > >to be preempted.
> > >
> > >Can you tell us what can cause these problems with invalid sessions?
> > >
> > >Thanks,
> > >
> > >Scott
> > >
> > >Condor Support Information:
> > >http://www.cs.wisc.edu/condor/condor-support/
> > >To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
> > >unsubscribe condor-users <your_email_address>
> >
> >
> > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> > Todd Tannenbaum                       University of Wisconsin-Madison
> > Condor Project Research               Department of Computer Sciences
> > tannenba@xxxxxxxxxxx                  1210 W. Dayton St. Rm #3357
> > http://www.cs.wisc.edu/~tannenba      Madison, WI 53706-1685
> > Phone: (608) 263-7132  FAX: (240) 359-5654
> >
> > Condor Support Information:
> > http://www.cs.wisc.edu/condor/condor-support/
> > To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
> > unsubscribe condor-users <your_email_address>
> Condor Support Information:
> http://www.cs.wisc.edu/condor/condor-support/
> To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
> unsubscribe condor-users <your_email_address>


-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Todd Tannenbaum                       University of Wisconsin-Madison
Condor Project Research               Department of Computer Sciences
tannenba@xxxxxxxxxxx                  1210 W. Dayton St. Rm #3357
http://www.cs.wisc.edu/~tannenba      Madison, WI 53706-1685
Phone: (608) 263-7132  FAX: (240) 359-5654 

Condor Support Information:
http://www.cs.wisc.edu/condor/condor-support/
To Unsubscribe, send mail to majordomo@xxxxxxxxxxx with
unsubscribe condor-users <your_email_address>