[molpro-user] h2o_status.com in quicktest hangs after completing

Benj FitzPatrick benjfitz at uchicago.edu
Mon Jan 30 15:36:57 GMT 2006


Nick,
I had some more time to test it out and here is what I found.
 When make quicktest is called I see 4 make processes spawn on
top which strikes me as odd (this is with patchlevel 518,525,
and 536)  Two test jobs are failing, and they have errors
where they can't kill the process.  The 2 jobs are sf6_dk and
, and here is a snippet of the output in the .errout files. 
This error keeps repeating.

------------------------------------------
 GA ERROR term_jobstep on processor   1

Killing pid 10924
Recieved sigusr1 in 10924
Attempting to die gracefully
MATION MAY BE LOST

 GA ERROR fehler on processor   1

Killing pid 10924

 RECURSIVE CALL OF FEHLER. EMERGENCY STOP.
 FILE INFORMATION MAY BE LOST

 GA ERROR fehler on processor   1
---------------------------------------

Thanks,
Benj FitzPatrick

---- Original message ----
>Date: Tue, 24 Jan 2006 11:57:27 +0000
>From: Nick Wilson <WilsonNT at Cardiff.ac.uk>  
>Subject: Re: [molpro-user] h2o_status.com in quicktest hangs
after completing  
>To: Benj FitzPatrick <benjfitz at uchicago.edu>
>
>Are all the testjobs passing up to this point?
>Nick
>
>Benj FitzPatrick wrote:
>> Nick,
>> I get a different problem now when I run quicktest.  At one of
>> the jobs (near h2o_direct, but I'm not sure exactly which one)
>> it spawns an additional pair of molpro commands so I have 4
>> processes going.  This bogs down the system tremendously so I
>> end up killing the jobs when this happens.  Then quicktest
>> goes on as before.  This happens with patchlevel 518 and for
>> multiple compilations of molpro (although they all use atlas).
>>  Any ideas?
>> Thanks,
>> Benj FitzPatrick
>> 
>> ---- Original message ----
>>> Date: Mon, 23 Jan 2006 19:23:21 +0000
>>> From: Nick Wilson <WilsonNT at Cardiff.ac.uk>  
>>> Subject: Re: [molpro-user] h2o_status.com in quicktest hangs
>> after completing  
>>> To: Benj FitzPatrick <benjfitz at uchicago.edu>
>>>
>>>
>>>
>>> Benj FitzPatrick wrote:
>>>> Is my patchlevel the # at the beginning of the PATCHLEVELS
>>>> file in the base directory?
>>> Yes it is
>>>
>>>   If so I have patchlevel 411 and
>>>> should probably get an updated source.
>>> You can get the latest patches by doing
>>>  make patch
>>> and then recompiling
>>>
>>> Nick



More information about the Molpro-user mailing list