April 16, 2012 at 8:37 am
sknox (4/16/2012)
call.copse (4/16/2012)
I would go further than the other posters on this thread - sorry, this question and the answer given are just plain WRONG.See this MSDN blog entry which clearly explains log sizes in different modes:
The LOG FILE BACKUP is certainly likely to be larger.
The LOG FILE (subsequent to backup) is almost certainly SMALLER given that it is not doing point in time restore.
I will accept that the question could have been mangled unintentionally in the editorial prcess or something - but, the fact remains, the correct answer to the question as written is NO.
I would say that since the question read "...could you see a log file bigger than usual?" the only logical answer is yes. While the bulk-logged recovery model is a feature designed to reduce log file growth, it does not constrain log file growth. Therefore you certainly could see a larger log file in some circumstances.
However, I would also say that that does not seem to be the intent of the question. The question was poorly worded, but the explanation is correct. So, poor wording of the question has given us a "If False then True" situation -- which, if my memory of basic logic serves me properly, is still True. So much for logic.
I disagree. There are no circumstances where immediately post log backup, a full recovery model database log file will be less big than a bulk logged model database log file after the same set of transactions. I am pretty sure there are few conceivable circumstances where it might even be the same size. The blog link I sent above explains this pretty well.
I'll echo Tom - I was pretty annoyed to spend time trying to get the question right according to the exact wording and find I still had it wrong - but I have not yet submitted any questions, so good on you. It must not be easy to put yourself before an audience who are by nature likely to be EXTREMELY pedantic.
April 16, 2012 at 8:53 am
I'll echo Tom - I was pretty annoyed to spend time trying to get the question right according to the exact wording and find I still had it wrong - but I have not yet submitted any questions, so good on you. It must not be easy to put yourself before an audience who are by nature likely to be EXTREMELY pedantic.
No problem, thank you for the feedback.
M&M
April 16, 2012 at 8:54 am
Thomas Abraham (4/16/2012)
Thank you for the question. I thought the question was straightforward, but then it's Monday.
I figured what was being asked, no confusion here either...
April 16, 2012 at 8:55 am
call.copse (4/16/2012)
I would go further than the other posters on this thread - sorry, this question and the answer given are just plain WRONG.See this MSDN blog entry which clearly explains log sizes in different modes:
The LOG FILE BACKUP is certainly likely to be larger.
The LOG FILE (subsequent to backup) is almost certainly SMALLER given that it is not doing point in time restore.
I will accept that the question could have been mangled unintentionally in the editorial prcess or something - but, the fact remains, the correct answer to the question as written is NO.
Tks for the question, however, I am another who answered based on log file and not log file backup... :pinch:
April 16, 2012 at 9:12 am
I was also thrown off my the 'log file' at the end of the question. It should have been 'could you see a log backup file larger than usual' By Log File I thought the question meant 'transaction log file' , which immediately after a backup either in Full or Bulk-logged should be the same.
Thanks
April 16, 2012 at 9:23 am
I got it wrong for legitimate reasons as I assumed the question was referring to log backups and not log files, and thought because bulk recovery model minimally logs bulk operations that both the log file and log backup file would be smaller so, from that angle, thanks for the question as I learned something new. However, in rereading the question I can certainly understand why some people would've received a different interpretation thus thought the question was asking about the log file and not the log backup file, but it's not as though this was the first poorly worded question we've had... 🙂 that being said, I WANT MY POINT BACK, HULK SMASH
April 16, 2012 at 9:35 am
call.copse (4/16/2012)
sknox (4/16/2012)
call.copse (4/16/2012)
I would go further than the other posters on this thread - sorry, this question and the answer given are just plain WRONG.See this MSDN blog entry which clearly explains log sizes in different modes:
The LOG FILE BACKUP is certainly likely to be larger.
The LOG FILE (subsequent to backup) is almost certainly SMALLER given that it is not doing point in time restore.
I will accept that the question could have been mangled unintentionally in the editorial prcess or something - but, the fact remains, the correct answer to the question as written is NO.
I would say that since the question read "...could you see a log file bigger than usual?" the only logical answer is yes. While the bulk-logged recovery model is a feature designed to reduce log file growth, it does not constrain log file growth. Therefore you certainly could see a larger log file in some circumstances.
However, I would also say that that does not seem to be the intent of the question. The question was poorly worded, but the explanation is correct. So, poor wording of the question has given us a "If False then True" situation -- which, if my memory of basic logic serves me properly, is still True. So much for logic.
I disagree. There are no circumstances where immediately post log backup, a full recovery model database log file will be less big than a bulk logged model database log file after the same set of transactions. I am pretty sure there are few conceivable circumstances where it might even be the same size. The blog link I sent above explains this pretty well.
The question did not ask if the log file would be bigger that a log file generated from a full recovery model database with the same set of transactions. The question asked if the log file could be bigger than usual. In this context, the logical interpretation of "usual" would be the average log file size preceding the model change. A single unusual non-bulk-logged transaction could account for that.
April 16, 2012 at 9:56 am
I guessed right, on the basis that if the answer was No then it wouldn't be a very interesting question 🙂
It then prompted me to read up a bit on bulk logging, so from my perspective was a good question.
April 17, 2012 at 12:57 am
Got it wrong, because I didn't know log file and log backup apparently are the same thing.
Anyhow, it made me think, so a part of the question was succesfull.
Need an answer? No, you need a question
My blog at https://sqlkover.com.
MCSE Business Intelligence - Microsoft Data Platform MVP
April 17, 2012 at 2:32 am
sknox (4/16/2012) The question did not ask if the log file would be bigger that a log file generated from a full recovery model database with the same set of transactions. The question asked if the log file could be bigger than usual. In this context, the logical interpretation of "usual" would be the average log file size preceding the model change. A single unusual non-bulk-logged transaction could account for that.
True enough - if you are getting away from the intent to cover the backup issue I think the wording is loose enough to accept that interpretation - although if you accept that interpretation then you have to say full recovery model vs bulk is a complete red herring and has nothing to do with the actual question or answer. The question is then simply 'could your log be bigger than usual?' which is pretty pointless, which is why I took the interpretation 'would the log file be bigger than a log file generated from a full recovery model database with the same set of transactions'.
April 17, 2012 at 7:33 am
sknox (4/16/2012)
I would say that since the question read "...could you see a log file bigger than usual?" the only logical answer is yes. While the bulk-logged recovery model is a feature designed to reduce log file growth, it does not constrain log file growth. Therefore you certainly could see a larger log file in some circumstances.However, I would also say that that does not seem to be the intent of the question. The question was poorly worded, but the explanation is correct. So, poor wording of the question has given us a "If False then True" situation -- which, if my memory of basic logic serves me properly, is still True. So much for logic.
This is pretty much the way I looked at this question. The key word to me was could as well, which pretty much forces the answer to be yes because it is such an open-ended question. I'm thinking along the lines of Arthur C. Clarke's first law:
When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
April 17, 2012 at 12:26 pm
Ernie Schlangen (4/17/2012)
sknox (4/16/2012)
I would say that since the question read "...could you see a log file bigger than usual?" the only logical answer is yes. While the bulk-logged recovery model is a feature designed to reduce log file growth, it does not constrain log file growth. Therefore you certainly could see a larger log file in some circumstances.However, I would also say that that does not seem to be the intent of the question. The question was poorly worded, but the explanation is correct. So, poor wording of the question has given us a "If False then True" situation -- which, if my memory of basic logic serves me properly, is still True. So much for logic.
This is pretty much the way I looked at this question. The key word to me was could as well, which pretty much forces the answer to be yes because it is such an open-ended question. I'm thinking along the lines of Arthur C. Clarke's first law:
When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
The trouble is that to many people (including me) it looks as if the question means "is it at all likely that the decision to switch to bulk mode increases the size of" whatever it refers to. If it refers to the log file (which is what it says) the answer is clearly NO. If it refers to the backup file (which is what the question author says he intended) the answer is clearly YES.
And yes, regardless of Clark's law I will state that it is extremely unlikely that a switch to bulk logging will cause an increase in the log file size, and that it's equally unlikely that it won't cause an increase in the log backup size. If you believe Clarke's law you will probably want to claim that when I (an elderly scientist - on whether distinguished or not I make no comment:-D) say that it is impossible that switching to bulk logged mode then doing a bulk operation then switching back to full mode and then doing a backup will lead to a smaller backup file that doing all the same except that the recovery mode switches are omitted I am likely to be talking from the wrong orifice, and that probable mistake on your part just goes to show that Clarke's law is (exactly as he intended) a broad generalisation that can't be used to determine specific outcomes although it can be used to indcate that unvalidated theories (as opposed to empirical knowledge) are often a waste of space.
Edit: eliminate double negatives (Cha toigh leam a' bheurla shasunnach - I hate the English gibberish).
Tom
April 17, 2012 at 1:19 pm
L' Eomot Inversé (4/17/2012)
Ernie Schlangen (4/17/2012)
sknox (4/16/2012)
I would say that since the question read "...could you see a log file bigger than usual?" the only logical answer is yes. While the bulk-logged recovery model is a feature designed to reduce log file growth, it does not constrain log file growth. Therefore you certainly could see a larger log file in some circumstances.However, I would also say that that does not seem to be the intent of the question. The question was poorly worded, but the explanation is correct. So, poor wording of the question has given us a "If False then True" situation -- which, if my memory of basic logic serves me properly, is still True. So much for logic.
This is pretty much the way I looked at this question. The key word to me was could as well, which pretty much forces the answer to be yes because it is such an open-ended question. I'm thinking along the lines of Arthur C. Clarke's first law:
When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
The trouble is that to many people (including me) it looks as if the question means "is it at all likely that the decision to switch to bulk mode increases the size of" whatever it refers to. If it refers to the log file (which is what it says) the answer is clearly NO. If it refers to the backup file (which is what the question author says he intended) the answer is clearly YES.
And yes, regardless of Clark's law I will state that it is extremely unlikely that a switch to bulk logging will cause an increase in the log file size, and that it's equally unlikely that it won't cause an increase in the log backup size. If you believe Clarke's law you will probably want to claim that when I (an elderly scientist - on whether distinguished or not I make no comment:-D) say that it is impossible that switching to bulk logged mode then doing a bulk operation then switching back to full mode and then doing a backup will lead to a smaller backup file that doing all the same except that the recovery mode switches are omitted I am likely to be talking from the wrong orifice, and that probable mistake on your part just goes to show that Clarke's law is (exactly as he intended) a broad generalisation that can't be used to determine specific outcomes although it can be used to indcate that unvalidated theories (as opposed to empirical knowledge) are often a waste of space.
Edit: eliminate double negatives (Cha toigh leam a' bheurla shasunnach - I hate the English gibberish).
I understand (and agree with) your point about the unlikeliness of the log file increasing in size. However, if it is possible for the log file size to increase, then the answer is "yes" to the question as stated (in my opinion, of course). And, for the record, I consider you to be "distinguished". I've certainly learned some things just from following up on references that you've made.
April 17, 2012 at 3:00 pm
Ernie Schlangen (4/17/2012)
I understand (and agree with) your point about the unlikeliness of the log file increasing in size. However, if it is possible for the log file size to increase, then the answer is "yes" to the question as stated (in my opinion, of course).
Yes, on a literal interpretation of the words the answer is "yes"; the trouble with that interpretation is we would be reading the question as "is it false that selecting the bulk recovery model guarantee that it is impossible for the log file to grow" and given the words chosen to pose the question is looked very unlikely that that was the author's intention (and indeed he has said it was not his intention - he meant to ask about the backup file, not the log file).
And, for the record, I consider you to be "distinguished".
Oh dear, something else we disagree on.
I've certainly learned some things just from following up on references that you've made.
thanks for telling me that - I'm glad to know that someone is following up the links/references I provide.
Tom
April 23, 2012 at 8:18 pm
Good question. Thanks for submitting.
http://brittcluff.blogspot.com/
Viewing 15 posts - 16 through 30 (of 32 total)
You must be logged in to reply to this topic. Login to reply