Re: OT Discussion- Priority of Performance Tuning...
Date: Sun, 30 Oct 2011 08:07:28 -0700 (PDT)
So I have to agree with this "Martin" guy, too? Hi Martin! :D Actually, considering all the responses, I think I have worked in at least one environment that has fit each of the situations, (and wouldn't be surprised if others have similar situations as well... :)) I have worked in an environment where I matched Wolfgang's view almost to the letter, but I am currently in an environment that does not and I am expected to be more pro-active, think more on my feet when it comes to tuning.
Example from the last two weeks:
Large release to production that changes large area of the design on how we process some data in daily processing. Developer in return, has to make changes to a development batch process that only runs on the weekend, is difficult to truly test fully because of resource power required or tuning time allocated to correct the original design, but does the best he can and moves it to production to run last weekend for the first time.
I pulled the deadlock trace files, contacted the developer in charge of the task, offered to go through them, run a few reports and tune without impacting any of my priority tasks. The business is unaware there is any issue with this important weekend process and I and the developer work together. I make the first pass as the main "issues" that I see and pass the changes to him to put in for this weekend's run. The business is again, not privy to any of this, as I offered my assistance to the developer, as any of the DBA's in our group would have and work through the task outside of our normal work load, (many know, I'm managing and am the main resource for upgrading our entire environment to 11g in a short timeline, along with the em12c project...)
This is where I must be aware of the importance of this process is to revenue, the pressure on the developer and make an executive decision, juggling my priorities and hours to address. Now other tuning work I would like to do is not business impacting, are just "it would be nice to have" opportunities and I will get to those when my other priorities are complete or when I have lull's in demands.
I'm hoping by giving examples that might make a bit more sense... :)
Sr. Database Administrator and Developer DBAKevlar.com
From: Martin Berger <martin.a.berger_at_gmail.com> To: kellyn.potvin_at_ymail.com
Cc: "breitliw_at_centrexcc.com" <breitliw_at_centrexcc.com>; oracle-l List <oracle-l_at_freelists.org> Sent: Sunday, October 30, 2011 3:55 AM
Subject: Re: OT Discussion- Priority of Performance Tuning...
after scanning through all the emails so far, I'd like to throw in some ideas myself.
From my point of view, yes, it's politics most of the time. I assume
this is not enough answer for you, probably you want to know the 'why'
So let's start with politics at all: What is this politics good for? In my simple mind, most of the politics ist the talking before a decision to get a clear, agreed picture about a situation. And often it's also the talking after the decision, to find the previous misunderstandings and blind spots to - lead to a new decision. Nothing bad about that for me. Maybe you don't like the amount of 90%, but that's kind of the culture of a company. As a bad example: some countries tried with much less talking & decision: one big (set of) decision was enough for the next 5 years. Do you remember the outcome?
Wolfgang and you discussed whether only the needs of business should
be delivered, or you should find your own point of view for the wealth
of the company and follow that road, too. Well, you both are right
(and therefore not really, at the same time):
There is only a little gap between Wolfgang and you: Wolfgang expects
the "business" to state all their needs explicite, whereas you accept
some needs implicite. (Even Wolfgang is not consequent with his
mentioning of backup/restore)
With this difference in mind we can come back to your question about politics, and now it's only about the flavor you prefer: Either talk about every little bit in advance, define it, refine even the first definitions until both sides have no space fir any (mis-)interpretation. Or let "business" just define their major golas and do al the politics later with dozens of '... I assumed ...' and '... but yo did not said that earlier ...' To make the situation worse, you both have your background to justify your answer: You where hired to do exactly this job, knowing what comes up before it's so worse it really affects business. Wolfgang seems to be more project based: there requirements should be well defined at the beginning, so everyone can agree about scope, timeline, prices and such things.
Based on most of the replies here, I assume, most of us live in a reactive world, where we have to start with little requirements but 'repair' a lot afterwards. None even mentioned my favorite topic yet: capacity planning.
To answer your question about my business, currently I'm working in a
mobile telco company in Austria.
For some reasons we can 'solve' nearly all issues in hardware, so the high need for getting changes into production leads to a mostly reaktive situation. Most of my time I spend doing politics to fix 'issues' as early as they raise (sometimes even earlier) and to keep this politics away from my team, so they can do the technical work. This is my way to tune the performance of my team ;-)
maybe you agree with these analysis,
On Wed, Oct 26, 2011 at 23:29, Kellyn Pot'vin <kellyn.potvin_at_ymail.com> wrote:
> I'll share my view on this one as I did with Wolfgang, as I still disagree with a bit of this, finding that I can't follow the business blindly, because sometimes doing what's best for the business isn't always doing exactly what they think should be done.
> I know how many times, after hours, I worked on something that I knew was going to rear it's ugly head and sure enough, my manager and the business would approach me, and I was prepared with what appeared to be a quick fix because I could see the impact to business even if the business could not. The business hired me to be that expert, do what they ask, but to also go that extra mile to "watch their backs". Yes, two different companies even referred to me as the Psychic DBA as I seemed to know what I needed to work on before they saw the importance of it. Does that mean I didn't take care of their priorities? No... It simply meant I had the foresight to keep on my radar what had the potential to go wrong and be pro-active about it. I am part of the business, an expert in my field and sometimes its my job to get the business what it needs even if it isn't what it says it was what it wanted...
> Kellyn Pot'Vin
> Sr. Database Administrator and Developer
> From: Wolfgang Breitling <breitliw_at_centrexcc.com>
> To: oracle-l List <oracle-l_at_freelists.org>
> Sent: Wednesday, October 26, 2011 3:19 PM
> Subject: Re: OT Discussion- Priority of Performance Tuning...
> I think this thread started because I took exception to the comment "priorities for the business may not be the same as a priority for the DBA’s". IMO they should be the same, i.e. the DBA should not make something a priority if it is not a priority for the business, has their backing. The only exception to that rule is when it comes to backups. That must be a priority for the DBA even if the business does not understand the priority of it - although they need to be reminded of it at every opportunity. With backups the priority would become obvious to the business only when it is too late.
> I'm curious about the opinions on that. Not the last bit about backups. I hope we do not have any disagreement there.