May 14, 2015 at 7:16 pm
CodeGuys (5/14/2015)
Please read my website for Interview Questions and Answers,1. SQL Interview Questions and Answers
2. Basics Interview Questions and Answers[/url]
3. MSSQL Interview Questions and Answers[/url]
I really appreciate you trying to help people but you've got a lot wrong and a lot that's "inaccurate". For example, if a DBA or SQL Developer actually gave me the answer you posted for the following question (both quoted below), I'd seriously consider ending the interview and escorting the candidate to the door even if they aced all the other questions.
10. What’s the difference between a clustered and a non-clustered index?
A clustered index directly affects the way tabled data is stored on a specific disk. This means that when a clustered index is used, data is stored in sequential rows based on the index column value. This is why a table can only contain a single clustered index. Non-clustered indexes directly affect the way physical data is stored and managed within SQL Server.
There are other inaccuracies and things that are wrong. If you're really are looking to help someone, do what anyone should do before an interview and read up in the actual documentation for the product instead of the inaccuracies of posts on the internet. Then post the correct answers on your blog. And I don't care how well known they are, what their certs are, or whether they're MVPs or not... even the best of the heavy hitters get things wrong. As a great man once said, "Trust but verify".
--Jeff Moden
Change is inevitable... Change for the better is not.
May 15, 2015 at 1:55 am
Enya? When the database is down? Surely it has to be Motorhead?
May 15, 2015 at 2:15 am
Chris Wooding (5/15/2015)
Enya? When the database is down? Surely it has to be Motorhead?
Definitely Enya. When the situation is very stressful, the last thing I want is music that's going to get me more hyped up.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
May 15, 2015 at 5:20 am
Hi Jeff,
Sorry for the inconvenience...Thanks for your valuable feedback...We will rectify the issues and update soon...
Thank you so much.
-------------------------------------------------------------------------------------
CodeGuys,
Check my website for frequently asked interview questions and answers.
http://www.techieclues.comMay 15, 2015 at 9:05 am
Jeff Moden (5/14/2015)
CodeGuys (5/14/2015)
Please read my website for Interview Questions and Answers,1. SQL Interview Questions and Answers
2. Basics Interview Questions and Answers[/url]
3. MSSQL Interview Questions and Answers[/url]
I really appreciate you trying to help people but you've got a lot wrong and a lot that's "inaccurate". For example, if a DBA or SQL Developer actually gave me the answer you posted for the following question (both quoted below), I'd seriously consider ending the interview and escorting the candidate to the door even if they aced all the other questions.
10. What’s the difference between a clustered and a non-clustered index?
A clustered index directly affects the way tabled data is stored on a specific disk. This means that when a clustered index is used, data is stored in sequential rows based on the index column value. This is why a table can only contain a single clustered index. Non-clustered indexes directly affect the way physical data is stored and managed within SQL Server.
There are other inaccuracies and things that are wrong. If you're really are looking to help someone, do what anyone should do before an interview and read up in the actual documentation for the product instead of the inaccuracies of posts on the internet. Then post the correct answers on your blog. And I don't care how well known they are, what their certs are, or whether they're MVPs or not... even the best of the heavy hitters get things wrong. As a great man once said, "Trust but verify".
+1
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
May 15, 2015 at 9:20 am
Nicholas Cain (4/17/2009)
Here's my personal favourite to spring:It's 3pm on a Friday, your production OLTP environment is down, every hour down costs the company one million dollars. People are phoning constantly, folks are rushing the hallways, running into your work area and freaking out wanting to know what you are doing, and why it's not fixed yet.
What do you do?
That separates the men from the boys (or women from the girls)
Step 0: Ignore panicked people, rely on super-geek-introversion to do this effortlessly.
Step 1: Ignore any suggested diagnosis that was given in the initial report and focus on reported symptoms only. Figure out WHICH component of the OLTP system is ACTUALLY down, which may not be the piece that appears to be down.
Step 2: Once the faulty component is found (may or may not be the database), get the RIGHT PERSON to work on it.
If you get the right person to work on the right component, that's half the battle.
October 4, 2016 at 9:43 am
Hi
I am going to write my solution to this, however some of it you should give as an interview question, some you shouldn't
The answer to this question is both a technical and social
Firstly - Keep Calm, I personally have been in similar situations and seen them happen, the worst case (and one that I had to jump in) was when I saw a Junior DBA burst in to tears when one of the MD's of a company of around 3000 people was shouting at him. Point out in a calm, business professional way that they are delaying the ultimate starting of the business, do not get emotional. As a DBA you should of already developed a thick skin 😀
So now the specifics
1) Deal with the interruptions. This should be either your managers job to head people off at the pass or if you follow ITIL to raise a ticket and pass that ticket to an Incident Manager. The manager (either your own or the Incident) should have procedures for handling company wide disruptions like this, so they should be handling it. If you are the manager (poor you), email the company (and/or the people interested) saying you are dealing with the problem, then give a time of the next update.
If you still get people interrupting your solution refer them to the Manager owning the issue or tell them you are working on it and will give an update by XX:XX Time (the time in the email), and if they get really pushy explain the longer they talk to you the longer it will take for a solution, if they get really pushy then record the time it took before they left you to get on with the work, and state that in any reports.
At this point you should also screen your calls and only answer calls from the Manager in charge of the issue, the exception is if the MD of the company comes by, then all bets are off 😀
2) Escalation
In my 16 or so years of experience as a DBA, I have seen a number of these issues, the majority has been to do with Hardware, however the Database tends to be the one to blame.
So I do the following
1) Attempt to log onto the Server with the Database, if I cannot, Escalate it to the Server Team
2) If I can log onto the Server then quickly open up perfmon and check resources such as disk utilisation
at the same time
3) Attempt to connect onto the SQL Server while on the server, is your database up ?, what does sp_who2 say? is there blocking/deadlocking what is causing it ? Are all the users who should be connected, connected (i.e. if you have a web front end is the user in the list ?, is your database suspect ?, is there disk space available ?
Basically what you do here is check anything that can stop the running of a SQL Server Database, the list here is too long to post here.
3) You have found the problem what next?
Sometimes it is not as easy as simply finding the issue.
Take the following case
The issue is due to the raid controller of the server going, which has caused corruption of the data, the last backup table (Transaction log) we 5 minutes before the event.
So it is up to pass this information asap AFTER you have complete proof of this, and come up with a resolution for it. Mine would be to move the database to another server and repoint all the connections to this server, unless of course your disk engineers can promise that no further corruption can take place, or they can fit new disks.
Again the key here is communication. Whatever the resolution is on such a big issue you will be one (a vitally important one) cog, so it is up to you to keep the person managing the issue updated with both the problem, and what you want to do about it. It should be the person Managing the issue who arranges the extra resources (i.e. people, servers ect) however it will be based on what you want to do to solve it.
So say if your database is a huge one and will take x hours to restore, you need to communicate that.
It is up to you as the DBA for a Database Issue to explain the Risks involved in a resolution, and what are going to do to mitigate those risks, and how long it will take to for your part of the resolution to take place.
Once your part of the resolution is completed (or during if you have the time) document what you did, when you did it, and how it took.
4) (Probably not one for the interview) I call this one 'The Noble art of *** Covering'
With such a large problem, there will be a meeting after to discuss it, called the Root Cause Analysis. This will have a number of people involved and someone will always ask 'Why did this take so long ?'
At this point the manager should chip in if they don't then it will be up to you (if it is a database issue) to justify what went wrong and what you did. How I have handled this in the past, is to give them a time line.
You will get a lot of credit if you were the first to find the issue that turned out not to be a database issue. However if it is database issue then you will need to explain the issue and then describe the short and long term solution you will/have put in so this problem either will never happen again, or will log the issue in large glaring letters somewhere where you find it quicker.
So in summary
1. Keep Calm (most important thing of all)
2. Communication is key - keep people informed of the issue, what you are currently doing to resolve the issue, and what you have done to resolve the issue
3. Do not be afraid to escalate if you need to
4. Document what you have done
Hope this helps and please don't be too harsh on my first post 😉
Peter
October 4, 2016 at 9:45 am
See below (wrong post) 🙂
Viewing 8 posts - 91 through 97 (of 97 total)
You must be logged in to reply to this topic. Login to reply