Bas,
Sorry if I did not catch in previous emails what the objective of this
patch. Would you mind explaining the circumstances that initiated the
need and the intent for such a patch (feature)?
For example, we have noticed with the various versions of MAUI that we
have used over the years that there available resources provided by
TORQUE to execute jobs but sometimes, it takes MAUI awhile to recognize
this fact. Is the feature implemented by this patch intended to fix
this problem?
Thanks.
Stewart
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Bas van der
Vlies
Sent: Thursday, August 27, 2009 2:45 AM
To: [email protected]; [email protected]
Subject: Re: [Mauiusers] [torqueusers] Re: maui + torque job start rate
Here is the patch. This patch add a new keyword to maui.cfg:
- PBSASYNCJOBSTART
Default value is FALSE. This means old behaviour pbs_runjob() is used.
If TRUE then pbs_asyrunjob() is used. This function start jobs faster
then
pbs_runjob()
Can this patch be applied to the maui source?
There is also a a new release for maui_2_deb with this patch added:
ftp://ftp.sara.nl/pub/outgoing/maui_2_deb.tar.gz
Bas van der Vlies wrote:
> Thanks for the info. I have finished the patch and tomorrow i will do
> some testing before i will post it
>
>
> On 26 aug 2009, at 19:35, Tom Rudwick wrote:
>
>> That sounds correct to me. That is what I have in my patch to switch
>> the behavior at compile time. We've been using it successfully for a
>> long time.
>>
>> Tom
>>
>>
>> Bas van der Vlies wrote:
>>> I have some time to implement the pbs_asynrunjob patch as option for
>>> maui.cfg. If i read this thread correctly i can safely remove both
>>> MPBSJobModify() that use 'neednodes' and change the pbs_runjob to
>>> use HostList as parameter instead of MasterHost.
>>>
>>> I want to make it configurable which function is used:
>>> - PBSASYNCJOBSTART
>>>
>>> default: FALSE (old behaviour)
>>>
>>> Regards
>>>
>>>
>>>
>>> Garrick Staples wrote:
>>>
>>>> On Thu, Apr 09, 2009 at 11:10:25AM -0600, Josh Butikofer alleged:
>>>>
>>>>> Actually, I just checked out the Maui source code and it looks
>>>>> like you will need to keep at least one of the neednodes calls
>>>>> (the one before the call to pbs_runjob()), as Maui is not passing
>>>>> a host list into pbs_runjob(). If Maui does pass in the hostlist
>>>>> to pbs_runjob(), the neednodes calls are probably not needed.
>>>>>
>>>> That's correct. If Maui passes in the hostlist, don't modify
>>>> neednodes.
>>>>
>>>> I gave up trying to fix this behaviour in maui years ago. Here is
>>>> the patch I've had in my own maui for a very long time.
>>>>
>>>>
>>>> Index: src/moab/MPBSI.c
>>>> ===================================================================
>>>> RCS file: /usr/local/nfs/src/cvs_repository/maui/src/moab/MPBSI.c,v
>>>> retrieving revision 1.14
>>>> diff -u -r1.14 MPBSI.c
>>>> --- src/moab/MPBSI.c 5 Nov 2005 02:42:08 -0000 1.14
>>>> +++ src/moab/MPBSI.c 23 May 2006 01:50:11 -0000
>>>> @@ -1792,6 +1792,7 @@
>>>> return(FAILURE);
>>>> }
>>>>
>>>> +/*
>>>> if (MPBSJobModify(
>>>> J,
>>>> R,
>>>> @@ -1826,6 +1827,7 @@
>>>> J->Name,
>>>> HostList);
>>>> }
>>>> +*/
>>>> }
>>>> else
>>>> {
>>>> @@ -1904,7 +1906,7 @@
>>>>
>>>> MJobGetName(J,NULL,R,tmpJobName,sizeof(tmpJobName),mjnRMName);
>>>>
>>>> - rc = pbs_runjob(R->U.PBS.ServerSD,tmpJobName,MasterHost,NULL);
>>>> + rc = pbs_runjob(R->U.PBS.ServerSD,tmpJobName,HostList,NULL);
>>>>
>>>> if (rc != 0)
>>>> {
>>>> @@ -1928,6 +1930,7 @@
>>>> JobStartFailed = TRUE;
>>>> }
>>>>
>>>> +/*
>>>> if (J->NeedNodes != NULL)
>>>> {
>>>> if (MPBSJobModify(
>>>> @@ -1949,6 +1952,7 @@
>>>> J->NeedNodes);
>>>> }
>>>> }
>>>> +*/
>>>>
>>>> if (JobStartFailed == TRUE)
>>>> {
>>>>
>>>>
>>>>
>>>
>>>
>> <ATT00001.txt>
>
> --
> Bas van der Vlies
> [email protected]
>
>
>
> _______________________________________________
> mauiusers mailing list
> [email protected]
> http://www.supercluster.org/mailman/listinfo/mauiusers
--
********************************************************************
* Bas van der Vlies e-mail: [email protected] *
* SARA - Academic Computing Services Amsterdam, The Netherlands *
********************************************************************
_______________________________________________
mauiusers mailing list
[email protected]
http://www.supercluster.org/mailman/listinfo/mauiusers
