Re: 10046, Unaccounted for Time, Aix 5.3 -- How to Determine Process Pre-Emption

From: Cary Millsap <cary.millsap_at_method-r.com>
Date: Wed, 2 Dec 2009 21:45:26 -0600
Message-ID: <3a2a84fc0912021945u6a865906ge4460812e208bd96_at_mail.gmail.com>



Another book that any Solaris user should have in the library is this one<http://www.pearsonhighered.com/educator/product/Solaris-Performance-and-Tools-DTrace-and-MDB-Techniques-for-Solaris-10-and-OpenSolaris/9780131568198.page>. Richard McDougall is one of the authors. He is a scheduled speaker at the Hotsos
Symposium <http://www.hotsos.com/sym10.html> in Dallas this March 711 (as is Kerry), which *you should not miss*.

Cary Millsap
Method R Corporation
http://method-r.com

On Wed, Dec 2, 2009 at 8:47 PM, Kerry Osborne <kerry.osborne_at_enkitec.com>wrote:

> I'm not too familiar with aix but Solaris has several scheduling algorithms
> (Real Time, Fair Share, Time Share, etc...) These have different ranges of
> priorities. Do a man on ps. It probably has options to show priority and
> scheduling class (i.e. the algorithm). Mixing different algorithms in a
> single VM can cause problems.
>
> Here's the command in solaris that shows that info:
>
> ps -ef -o user,pid,project,class,zone,pset,pri,nlwp,psr,time,args
>
> As already mentioned, vmstats/sar should also give you a good clue as to
> whether processes are struggling to get on the CPU.
>
> There is a great Solris Internals book that describes all the scheduling
> algorithms among other things. I presume there is a similar reference for
> AIX. Also, I'd recommend Millsap's book for a better understanding of
> how/why unaccounted for time occurs.
>
> Kerry Osborne
> Enkitec
> blog: kerryosborne.oracle-guy.com
>
>
>
>
>
>
>
> On Dec 2, 2009, at 2:13 PM, Larry G. Elkins wrote:
>
> Listers,
>>
>> SA's asking for assistance on how to tell if a process is being impacted
>> and
>> to what degree by processor pre-emption. AIX 5.3. Asked in the context of
>> trouble-shooting some significant unaccounted for time issues in 10046
>> trace
>> files (e.g. 14 of 16 seconds unaccounted for vendor process, 300 of 700
>> seconds unaccounted for vendor batch job, etc). SA's have "ruled out"
>> overhead of writing the trace file so now looking at the pre-emption
>> aspect.
>> A bit out of my league, and I find plenty of approaches in IBM docs, and
>> from people on this list (but geared more towards Solaris and other
>> flavors,
>> not Aix). 9.2.0.8 EE and 10.2.0.4 EE, "static" and dynamic LPAR's, seen on
>> various boxes and databases.
>>
>> Larry G. Elkins
>>
>>
>> --
>> http://www.freelists.org/webpage/oracle-l
>>
>>
>>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Wed Dec 02 2009 - 21:45:26 CST

Original text of this message