October 2, 2011 at 1:38 pm
Comments posted to this topic are about the item Taking Risks
Brad M. McGehee
DBA
October 2, 2011 at 2:01 pm
Ah... sorry, Brad... I have to disagree with you on most of what you said except for "pushing a change to production without testing it first on a test server, should always be avoided".
First, I think that most DBA's that I know (and I know more than a hundred) do take the very risks that you speak of. The only thing they're not (and I'm not) willing to take a risk with is the safety of the data in a production database.
You want to upgrade a production box from 2005 to 2008 R2? Not a problem... build a parallel machine and work out all the bugs and faults first because experience tells me there's no such thing as a "transparent" migration. Then, do full regression testing of the apps because, sure as shootin', someone has embedded code somewhere and it can break during a rev change just like anything else (happend to us just about 3 months ago and our system DBA's had to migrate the data back to 2005.).
If you think all that's a bit cowardly, consider this... you know the risk of jumping out of an airplane with no parachute... so why don't you try it just based on someone's word or in the spirit of taking a risk? Heh... the answer, of course, is that you know better.
Same principle applies here... 😉 All good DBA's are secretly born in Missouri... the "Show Me" State. 😛
And, remember... they don't call it the "bleeding edge" for nothing. :hehe:
--Jeff Moden
Change is inevitable... Change for the better is not.
October 3, 2011 at 2:12 am
In terms of personal risk taking/development, I love my motorbike. So the last week I arranged an Enhanced Rider Scheme course for the following day.
Whilst I was REALLY nervous about what I was going to have to do and the new speed of risk assessment that I was going to have to pick up (remembering that not only is this life and death stuff... it's my own!), I thoroughly enjoyed myself and this morning on the way into work I felt a lot safer and cut through the traffic a lot smoother/quicker.
Therefore, I partly agree.
I do however agree that as long as the precautions don't detract from your aim, some risks are acceptable.
Life after all, is a balance of risk vs. benefit and it's up to us to decide the ratio we're happy with.
October 3, 2011 at 2:21 am
Try professional gambling for a living - that sure teaches you a thing or two about risk taking.
October 3, 2011 at 5:13 am
This coming Saturday i am having my first flying lesson. I have flown before but i have st myself the short term challenge of learning to fly and getting a private license within 18 months. And the then long term goal would be to get and maintain mmy own plain to fly the wife and kids around europe.
I volunteer at a RAF museum and look after the jets and bombers, so the mechanics are nothing new to me but learning to 'drive' them is. Its a natural progression really.
As for 'risk', well i guess there is risk involved but only if im really unlucky and a bit crap!
Adam Zacks-------------------------------------------Be Nice, Or Leave
October 3, 2011 at 5:16 am
You lucky thing Adam!
I thought a motorbike was exciting! 😛
Have fun
October 3, 2011 at 5:20 am
Ben Moorhouse (10/3/2011)
You lucky thing Adam!I thought a motorbike was exciting! 😛
Have fun
Thanks 🙂
CANNOT WAIT.
Im also learning from the World, UK, European Micro-Light champion..... He's my instructor. So i think im in safe hands.
When you start looking into ways to 'own' its not as pricey as people think. Not really much more then a small family car.
Adam Zacks-------------------------------------------Be Nice, Or Leave
October 3, 2011 at 5:23 am
Wow! Now that I hadn't expected!
Don't you have to pay for the airport etc?
October 3, 2011 at 5:32 am
No/ Micro-lights and light aircraft generally operate out of private airstrips or airfields, so all you really pay is hangerage at the airfield where you keep it. Generally speaking, most strips are free to land in (given prior notice).
Its not like with jets and the like that need long runways and facilities. In fact if you go anywhere near a large airport your liable to face heavy fines and police action.
Adam Zacks-------------------------------------------Be Nice, Or Leave
October 3, 2011 at 5:37 am
Ok... you good folks and "risk" takers that have posted so far... 😉
How many of you are actually willing to install a new product (or even "just" a switch from 2k5 to 2k8) on your production boxes without first testing it on a copy of the production or at least a subset-copy of production?
I hope the answer would be "none" but I have a feeling that's not going to be the answer.
--Jeff Moden
Change is inevitable... Change for the better is not.
October 3, 2011 at 5:41 am
I deal with fairly low numbers of users on most databases (~14,000 people) so I don't have a massive issue with diving straight in, because quite often the benefit of just doing it far outweighs the risk of it going wrong.
That being said, there are backups around should something hit the fan.
As I said - it's all about balancing risk with benefit, not completely ignoring it or being governed by it.
October 3, 2011 at 5:48 am
Heh... have you ever tried to restore a 2k8 backup back into a rebuilt 2005 server after a couple of days when you find out some of the lesser used features of your legacy apps don't work so well with 2k8? Tons'o'fun. 😉 Even on relatively "small" databases, it's just not worth taking the risk for.
But, to each their own. I think Brad's article was more of a "poll" anyway. 🙂
--Jeff Moden
Change is inevitable... Change for the better is not.
October 3, 2011 at 5:55 am
No offence, but who said anything about switching between versions?
In my mind, 2008 > 2005 isn't a proper backup.
I have 2005 backing up to a 2005 box and being restored and checked (granted very basic checks) daily.
If something were to go wrong, then I could just failover to the other box whilst the issue is fixed on the primary box.
October 3, 2011 at 6:09 am
Ben Moorhouse (10/3/2011)
No offence, but who said anything about switching between versions?In my mind, 2008 > 2005 isn't a proper backup.
I have 2005 backing up to a 2005 box and being restored and checked (granted very basic checks) daily.
If something were to go wrong, then I could just failover to the other box whilst the issue is fixed on the primary box.
Equally, no offense, but Brad did in his article...
Upgrading from SQL Server 2005 to SQL Server 2008 is a risk;
And you've made my very point... trying to a restore from 2k8 to 2k5 to undo a bad upgrade isn't possible. And just failing over to "the other box" would be a futile step even if it were a 2k5 box because, after collecting data on the new 2k8 box for a couple of days, the 2k5 "fail over" box would be quite out of date.
--Jeff Moden
Change is inevitable... Change for the better is not.
October 3, 2011 at 6:47 am
While risk taking is a psychological characteristic, many people including myself, avoid random risk taking, but will take selected risks (Steve might know about about the back road to Hurricane Pass where the wheels of my Jeep were literally inches from a couple thousand foot drop).
The same applies to systems we run. I am in NO hurry to move stuff off of one database platform that is working just fine onto another unless there is a direct benefit to doing so (like end of support). There are enough unpredictable variables in life, introducing extra ones withoug significant payback I think is counter productive.
...
-- FORTRAN manual for Xerox Computers --
Viewing 15 posts - 1 through 15 (of 27 total)
You must be logged in to reply to this topic. Login to reply