[torqueusers] Torque 2.4.4

Ken Nielson knielson at adaptivecomputing.com
Wed Jan 20 08:11:55 MST 2010


Joshua Bernstein wrote:
>
>
> Ken Nielson wrote:
>> Joshua Bernstein wrote:
>>> Glen Beane wrote:
>>>  
>>>> On Tue, Jan 19, 2010 at 5:01 AM, Douglas McNab 
>>>> <d.mcnab at physics.gla.ac.uk <mailto:d.mcnab at physics.gla.ac.uk>> wrote:
>>>>
>>>>     Hi,
>>>>
>>>>     I have been testing with 2.4.4 and it seems the OSC MPIEXEC Bug 
>>>> has
>>>>     crept back in with 2.4.4.     Could someone check and confirm 
>>>> this?  I have been running a simple
>>>>     MPICH testing and I have been getting:
>>>>
>>>>     /mpiexec: Error: get_hosts: pbs_statjob did not return 
>>>> "exec_host" //info.
>>>>     /
>>>>
>>>>     I thought this fix was in main after 2.4.3?
>>>>
>>>>
>>>>
>>>> I just checked subversion, and src/server/stat_job.c has not been 
>>>> modified since the 2.4.3 fix that was supposed to fix the bug 
>>>> preventing mpiexec from working properly.
>>>>     
>>>
>>> I see the same. It looks like somebody missed a merge over to the 
>>> 2.4.4 branch...
>>>
>>> -Josh
>>>
>>>   
>> The fix for this was checked in on December 2, 2009. The revision was 
>> 3268. It looks like the change is still in there. Attached is the 
>> diff file with the change. It was a matter of removing code that was 
>> making the call fail. This code is not in the 2.3-fixes branch.
> The code that was removed was not in the 2.3-fixes. Also there was no 
> bug in 2.3-fixes. It has been long enough that I can't remember what 
> the exact problem was but the code that crept into 2.4-fixes was the 
> problem. We need to probably do a debug session if we can and see just 
> where the call is failing.

Ken

> Maybe we have a new bug then that was created with 2.4.4? Any reason 
> why this wasn't pushed back to the 2.3-fixes?
>
> -Josh



More information about the torqueusers mailing list