« first day (1682 days earlier)      last day (3179 days later) » 
02:00 - 20:0020:00 - 23:00

2:05 AM
1
A: Is it possible to develop a SSIS with a SQL Server Evaluation Version

billinkcFor SQL Server 2005 and 2008 R2, the only way to get the Integration Services development environment, Business Intelligence Design Studio a.k.a. BIDS, was through the installation media aka you already bought Developer, Standard or Enterprise Edition. However, if you're looking to purchase SQL S...

Good answer.
 
ta
Been out teaching a class all day but there were two whole SSIS questions posted. I had a sad
 
@Phrancis I see you're in the room, hey why did you mark this answer as Community Wiki? Just curious. Answering your own question is encouraged.
@billinkc It's great when someone adds a much better answer, even where there's a perfectly ok accepted answer already.
 
2:21 AM
If I wasn't full of hatred over my client's flubbed Evaluation edition install, I'd have probably let it stand
 
@PaulWhite Must be out of habit, selfie-answers on Code Review are very often posted as CW, unless the OP has posted their own thorough Code Review with explanations and all (which happens but is fairly rare)
 
@Phrancis Well I think it's good you took the time to share the solution with the community. Nothing wrong in getting a bit of rep (and the green checkmark) for that. Would you like the CW removed?
 
@PaulWhite That would be good, thanks!
 
@Phrancis Done. Only you can mark it accepted (no rep earned for that though).
Sweet.
 
I thought I had done that too... must have been late that day
 
2:31 AM
 
@PaulWhite I see you just get voted in a moderator?
 
^^^ Code comments
 
@Phrancis Yes, bit keen at the moment. It'll wear off no doubt :)
 
Congratulations!
 
3:29 AM
this is the exact sort of crap that folks warned of when k-man went on his clustered index rant
2
Q: Should the PK on an identity column be non-clustered?

u23432534For a table with PK on an identity column, it will be clustered by default. Could it better be non-clustered? The reason is other indexes will be created for queries. A query which uses a non-clustered index and returned columns are not covered by the index will use less LIO because there is no ...

 
@swasheck Yep. Just thought it would be nice if we could have a single point where the pros and cons are listed. If the bounty expires with lack of interest, it'll probably end up closed for good. I might answer it myself though.
 
@PaulWhite i find it unanswerable for the reason(s) in my first comment
 
@swasheck Fair enough :)
 
but if there's anyone i'd trust to dazzle them with brilliance, it's you, @PaulWhite.
 
Suck up!
 
3:41 AM
i have to now that you're all blue and diamondy
 
No you should continue to treat me with the traditional level of contempt and horror.
 
i fear you've confused me with @billinkc. i try to treat you with fawning and awe
 
Oh yeah, right.
 
anyway ... i get the impression that the OP on that question may be misapplying the information in kejser's post. am i off-base there?
@PaulWhite i look forward to your answer
 
@swasheck No, you're probably right. K-man's points are valid in cases many people will not encounter.
 
 
2 hours later…
6:06 AM
@PaulWhite You have no choice. It says in the revision history that you are required to answer.
Maybe I go my prepositions mixed up :)
 
6:20 AM
@MikaelEriksson I don't think you did. It's probably the word order that adds the unnecessary ambiguity and spoils the good joke. Ideally, you would want it to read, "Canonical answer by Paul White♦ required" :)
 
@MikaelEriksson Ha ha ha awesome!
If I do answer, I'll award the bounty to someone else if possible.
 
There probably exist a special badge for a moderator-self-answering-bounty-awarding-action.
It could be the Pompous badge :)
 
6:35 AM
Ooo I hope so. I'd display that with pride!
 
6:48 AM
@PaulWhite dba.stackexchange.com/questions/110863/… - is this the start of a reign of tyrrany? :-)
 
@Vérace No, but there's really no point delaying the inevitable. I'd be as quick to correct an error or reopen.
And it's only "on hold". No answers, but comments and question edits are unaffected.
It certainly helped that your question was interesting, well formatted and clear, which puts it above the majority of first-time posts. But the social climate on dba is more cooperative than on the bigger sites even for less useful questions - which may contribute to better questions in turn... — Erwin Brandstetter 17 hours ago
It seems we are nice :)
 
Starts like a comment but ends like a proper answer.
 
Don't get me started on answers in comments :)
@AndriyM Re: this It's either a tutorial question (it which case it would be nice if he had a go), or a misunderstanding that CASE is somehow required to solve the problem. Personally, if I knew Oracle, I would just write a good way to get the required result as an answer and hope for the best.
 
7:05 AM
@PaulWhite It did read to me as a tutorial question.
@PaulWhite Incidentally, do psychic abilities come with a diamond?
I was going to ask you for your opinion about exactly that
 
For example in SQL Server:
DECLARE @Example AS table
(
    Eid integer NULL
);

INSERT @Example
    (Eid)
VALUES
    (1),
    (2),
    (3),
    (4),
    (5),
    (NULL),
    (NULL),
    (NULL),
    (NULL),
    (NULL);

SELECT
    [Number] = COUNT(E.Eid),
    [Null] = COUNT(*) - COUNT(E.Eid)
FROM @Example AS E;
And if the presentation is important, replace the final query with:
SELECT
    U.Eid,
    U.Counts
FROM
(
    SELECT
        [Number] = COUNT(E.Eid),
        [Null] = COUNT(*) - COUNT(E.Eid)
    FROM @Example AS E
) AS R
UNPIVOT
(
    Counts FOR Eid IN ([Number], [Null])
) AS U;
Ugh, sorry about the wall of code. I thought it would auto-limit.
 
Yes, that would impress the teacher all right :)
 
@AndriyM It wasn't really psychic. Sometimes you get a feel for how to do the translation from non-English speakers.
@AndriyM Probably not, since it's Oracle!
@AndriyM Well, I'd like to think we can pretty much always help if the requirement is discernable. I doubt the question has a long-term future though.
 
@PaulWhite I have no problem helping people learn simple techniques, but some things are too simple, to the point that I would just be reciting the manual. I just don't see much point in that.
I mean, nothing else but reciting the manual
The post in question could be either from a lazy student or from someone who can't put together the simple things they have learnt to solve a (relatively) complicated problem – that I'm perfectly fine with, hence my comment trying to establish what case that is.
 
@AndriyM Oh I agree.
 
7:18 AM
@PaulWhite There was a smiley in there - and the point about delaying is well made - From Marcel Pagnol - Jean de Florettte (my favourite author - my tranlsation) - "when you've started to strangle the cat, you have to finish it!"
 
ha ha nice!
 
7:35 AM
A mongodb expert we should encourage to participate.
 
not an answer or advanced spam? :)
 
@TomV Pre-spam nonsense. Flag it as spam or NAA and I'll destroy it and the user.
 
7:52 AM
i flagged it as NAA before you replied
 
Hm not seeing that flag.
 
If I click flag again i get: you have already raised this type of flag
 
Weird. OK.
Oh, NAA goes to the Low Quality Queue first for a while.
Deleted and destoyed, thanks @TomV
104
Q: Let's get rid of the 10K flag queue

Shog9The 10K tools are pretty cool... You get a birds-eye view of activity on the site, a "dashboard" view of what's happening. Some of the individual tools haven't scaled particularly well with Stack Overflow's growth, but the concept behind them is still sound: we trust you to enough to be a bee wat...

 
@PaulWhite You don't need to change much to make it work in Oracle:
 
@AndriyM How interesting! Thank you. I had no idea the syntaxes would be so close in this case.
 
8:05 AM
(SQL Fiddle doesn't work)
 
Hardly ever does these days.
Thanks it is helpul. But i need it in case conditions. — Babu 21 mins ago
LOL, of course.
 
"I promise I'll be a diligent student in future, just help me this time, pleeeeease", they might be thinking.
 
Maybe. I did a CASE thing, but I'm not spending any more time on it.
It's headed for closure pretty soon I think.
He's just not engaging.
@AndriyM Does that syntax translate to Oracle?
 
8:34 AM
@PaulWhite Yes, it's the same syntactical differences as before. Change the column alias style (to expr AS alias), remove AS from dataset aliasing – and it works in Oracle.
 
@PaulWhite ah ok, I didn't know spam flags get more attention than NAA
 
9:27 AM
 
9:38 AM
@swasheck I went off that Q when he said in one comment " the query I listed is the common one" and 5 minutes later he made another comment and edited the question to include more "common" cases.
 
9:49 AM
@Lamak Yes, thank you.
 
10:01 AM
@ypercube Thanks! All done. Sorry I was watching the 200th SG-1
This is more of a comment than an answer. Could you please flag it for moderators to change it into a comment. You can comment on all posts when you have enough reputation. P. S. Welcome to the forum! — Vérace 2 hours ago
Hey @Vérace . Couple of things about this ^^^
 
@PaulWhite You have the super ping privilege now. Super ping them!
 
As far as I can tell it is an attempt to answer the question. It might not be a very good answer (I don't know), but I wouldn't covert it to a comment.
@ypercube Well I don't want to wear it out :) Besides, doesn't super-ping forcibly drag people back to their computers, even if they're out of the house, or in the shower? I couldn't do that to him. ;)
So, poor answers: Down vote, comment to ask for more info & suggest improvements, maybe even flag as VLQ if you feel an answer is that bad, but don't ask the answerer to flag for a conversion to a comment if it shouldn't be.
86
A: Why do I need 50 reputation to comment? What can I do instead?

PëkkaWhy does this limit exist? We realize that new users may have valuable comments, and that we may lose some of those contributions by requiring 50 reputation points to unlock the feature. However, history and experience have shown that the downsides of allowing everyone to comment are far great...

(related)
So 99 times/100 if a user can't comment, and posts an answer instead I'll either (a) leave it because it is at least an attempt to (partially) answer; or (b) delete it.
That's my current thinking, anyway.
 
10:25 AM
I agree.
 
is relieved
 
10:45 AM
@PaulWhite "Couple of things about this ^^^" Is it the obvious spam that I missed? It's so rare that I wasn't even aware of it. Will keep an eye out in future.
Why is it when you put text between angle brackets (aka &gt - greater than symbol and &lt - less than symbol) that the text disappears? I've been caught by that a few times.
 
@Vérace No, it's a different answer.
 
-1
A: Why does "GRANT SELECT ... TO Role;" does not allow members of Role to SELECT?

user72700The role you created is password protected. Therefore it is not enabled by default, when alice logs in.

@Vérace That one. Your comment was:
41 mins ago, by Paul White
This is more of a comment than an answer. Could you please flag it for moderators to change it into a comment. You can comment on all posts when you have enough reputation. P. S. Welcome to the forum! — Vérace 2 hours ago
 
@PaulWhite Can you give me the URL to the question - can't seem to find it in my history. Also, what was the problem that you're finding?
 
It's the GRANT SELECT ... TO Role answer just up above. The comment has gone now.
You asked the answerer to flag for conversion to a comment.
Are you on a phone or something? Making it difficult to read stuff?
 
@PaulWhite I'm not sure I get you. What did I write and what was wrong with it? And no, I'm not on the phone - maybe tired - not sleeping so well. Was my error egregious?
 
11:02 AM
@Vérace This. No, not egregious. As far as I can tell it is an attempt to answer the question. It might not be a very good answer (I don't know), but it shouldn't be a comment.
The rest of the reasoning is just back up the Heap a page or so. Take a look when you've had some sleep :)
 
@PaulWhite OK - an attempt at an answer - even short and lacking in explanation, it doesn't become a comment. If it's NAA, then I can comment about flagging. Got it.
 
@Vérace If it's genuinely NAA, flag it as such.
...
Meanwhile:
No actually a interview quesion. they have asked paticularly to write the case statements to solve this. — Babu 7 mins ago
If it turns out he's asking the question while in an interview I will scream.
 
11:18 AM
Regardless of whether they are being interviewed just now or it was yesterday or a year ago – my only question is, "seriously?"
Not a real question, of course, just a reaction
 
I suppose it makes a change from Fizz Buzz :)
 
@PaulWhite "If it turns out he's asking the question while in an interview I will scream.". dba.se's (and StackOverflow's) reputation has travelled far and wide. I'm a mature student and we were specifically told not to use these sites even in open-book exams! :-)
 
@Vérace Goodness me, really? Wow.
 
11:52 AM
@PaulWhite Maybe I didn't express myself well. I mean, there's nothing stopping us using them day to day - just that for exams, they're not allowed, even for open-book exams. Kinda makes sense - just as IM isn't allowed or email. You are allowed Google though - so if Google leads one to dba.se - I think it's more the communication with 3rd parties during exams that they're on about, so you could read dba.se, but just not post during tests.
 
I understood you, just surprised Database Administrators is explicitly mentioned. Nice!
 
 
2 hours later…
Kin
1:28 PM
Morning / afternoon All
 
Morning
This weekend is going to be rough. #hot http://t.co/jpLunzF5u7
113/114F.
 
:)
the hottest weekend this year
 
That's ridiculous. I struggled with 99F last week.
 
JNK
well when you live in Arizona it's not like a big surprise right
 
@JNK it's still rare to be that hot
 
JNK
1:41 PM
wait are you in AZ or NM
 
@PaulWhite I would die immediately
 
JNK
I am doubting myself
 
@JNK Yeah but even so. Fahrenheit temperatures should not need three digits. Ever.
@Lamak Yep me too.
 
JNK
@PaulWhite well not in celsius at least
 
@PaulWhite we don't have the humidity of other places but we still have some - for example at 4:30am today it was 91F with 45% humidity
@JNK AZ
 
JNK
1:42 PM
ok thought so
@PaulWhite in Kelvin it definitely is OK
 
Oh shush!
 
JNK
oh man you're such a historical revisionist
 
No I'm not
:)
@ypercube Sorry I couldn't preserve your comment. It was indeed the same user.
 
JNK
@PaulWhite so do you have thoughts on picking a new flag?
I saw that on the news and thought of you since you're the only real person from NZ I am aware of
I think it's awesome some have ferns on them
 
Oh, ha ha, I thought you meant a new flag on the site at first.
 
1:51 PM
Monking
 
Yes I'm strongly in favour of selecting a new flag (though we might not reach consensus on which one, meaning we're stuck on the old one).
I like most of the "long list" designs.
I think we should avoid too much black, or one that looks too much like an existing sport flag.
 
@PaulWhite Do you find it odd that the optimizer knows how to simplify things like P.BusinessEntityID >= 1 AND (P.BusinessEntityID <= 64 OR P.BusinessEntityID < 20 OR P.BusinessEntityID < 5) to a single range predicate, but (apparently) doesn't do the same with something like WHERE P.BusinessEntityID IN (1,2,3,n,..,64)?
 
@JamesLupolt Yes and no.
 
@PaulWhite I suppose that's a good answer to a vague question. : )
 
One is domain simplification over a complete range, the other would require knowledge of discreteness as well. TF 2301 enabled some trickery in that area, but only for statistics, not q.o. per se.
 
1:58 PM
@PaulWhite - quick question; is it correct to state that rows in a heap that are deleted or truncated are not returned as free space as efficiently as rows that are removed from a clustered index?
 
@JNK That one came a bit out of left field, but my wife and I both quite like it.
 
@PaulWhite I must be misunderstanding something. The list (1,2,3,n,..,64) is not necessarily discrete?
 
good afternoon/evening/morning, by the way!
 
It is a nice flag.
 
isn't it 3am in NZ?
 
1:59 PM
@JamesLupolt To turn it into [1, 64] i.e. >=1 AND <= 64 you'd need to know that integers are always whole numbers.
@MaxVernon 02:00
@MaxVernon Good morning.
 
ok
lol. yes good morning!
 
@MaxVernon The main difference I am aware of is that empty page deallocation requires a TABLOCK hint (or equivalent) for DELETE, and even then it's not guaranteed in all circumstances. TRUNCATE TABLE does not suffer from the same issue IIRC.
That flag design is Unity Koru
 
I've had some fun issues recently caused by heaps full of empty space taking up most of a buffer pool.
 
we have a couple hundred staging tables that are heaps. We're seeing that space is not being freed in the database resulting in very large database sizes. Another DBA here is thinking converting those staging table heaps into clustered indexes would resolve that particular issue. So, we converted all those staging heaps into clustered indexes with what I thought was a pretty good key - QA is reporting load times are 3 times greater.
(all this happens in the test environment, of course)
 
@PaulWhite Thanks. Yes, I see your point now. I guess there would have to be some additional (and potentially tricky) reasoning about types there. And it might not be worth it.
 
2:06 PM
I'm thinking loading into a clustered index should be a small amount slower, but not three times slower.
 
JNK
@PaulWhite I like the idea of picking a new one
 
@JamesLupolt Nah it'd be totally worth it. They're just too lazy (LOL!)
 
@JNK Yeah we need to stop looking like Australia.
Spot the difference. (AUS top; NZ below)
 
JNK
I had no idea they were so similar
 
2:12 PM
@JNK So yeah I'm not going to bore you by listing every design I prefer, but the Kyle Lockwoood Silver Fern ones are very good; Land of the Long White Cloud by Mike Archer, Embrace by Denise Fung, and finally Manawa by Otis Frizzell.
@JNK It's embarrassing.
 
I didn't know there were torx screws on the New Zealand flag
3
 
/chatban is available :)
The four others I like.
24 mins ago, by Paul White
user image
There ya go.
Bill will ask where the kangaroo design is, of course.
 
2:30 PM
@PaulWhite you need more hobbits and rings on the flag
 
Indeed. And sheep, no doubt :-|
 
your words, not mine
 
Kin
@PaulWhite Looks nice
 
@Kin Which one? There are four potential designs there. Well five if you count the other one :)
 
Kin
@PaulWhite Yeh .. especially the the feather one
 
2:37 PM
@Kin Oh you mean the bottom one? Yes we like that one best at the moment. It's a Koru - an unfurling fern leaf - though, not a feather.
Huh, well here's one I never expected to earn:
Perhaps that's an error. I'm no Generalist!
 
yeah. you're more of a kernel
 
@PaulWhite I guess half of the top tags is either performance or sql-server-versionnumber
 
@TomV yeah. Also tags like "index", "trigger"
 
Looks that way yes
Still, only 17 awarded, so that's pretty collectible. Could be worth $$$ one day.
 
3:02 PM
0
Q: how to pass the "file path" as a parameter in powershell?

marcelo miorelliI have recently added a SQL Server JOB on SERVER-A that goes to SERVER-B and backups all the SSAS databases. This seems to be fine. However when I want to delete backups older than 3 days, using the function below, I get the error message on the picture below. It says that the path is wrong,...

not sure how that is a dba question
maybe it turned into one with the access denied
 
3:16 PM
@TomV powershell is a DBA thing
just ask @MikeFal :)
 
Powershell is very much a DBA thing
 
Just not a DB-Dev thing :)
 
Get-DropPowershell
 
3:38 PM
@Marian yes but the original question was just about deleting files that happened to Be backups
 
Kin
@PaulWhite Nice .. Koru ...symbolizing new life, growth, strength and peace (en.wikipedia.org/wiki/Koru)
 
Yep.
 
Kin
4:40 PM
dba.stackexchange.com/q/110945/8783 seem a too broad topic
VTC please ?
 
-- This print raises an error:
-- Conversion failed when converting the varchar value 'An error was raised\nSEVERITY: ' to data type int.
PRINT 'An error was raised'
    + '\nSEVERITY: ' + @ErrSeverity
    + '\nSTATE: ' + @ErrState
    + '\nMESSAGE: ' + @ErrMsg
    ;
Anyone got a clue what I'm doing wrong?
 
yes, some of those variables are int, therefore you can't directly use + to concatenate them
 
@Phrancis You are not converting a number to a string
 
Oh, gotta cast them eh, good catch
 
yes, the implicit conversion direction is char -> int, not the other way
 
4:52 PM
Superb, thanks :)
 
@ypercube so, was the euclid game too easy for you?
 
@Lamak I've played it before.
But thank you for losing 10 more minutes today.
 
yeah, me too, last year. But then I lost the link and finally found it again
 
GeoGebra is a nice tool
 
yup, very nice
 
5:17 PM
@Kin Way off topic. Burned.
 
5:28 PM
@Phrancis do this instead:
PRINT 'An error was raised'
    + '\nSEVERITY: ' + CONVERT(NVARCHAR(2048), @ErrSeverity)
    + '\nSTATE: ' + CONVERT(NVARCHAR(2048), @ErrState)
    + '\nMESSAGE: ' + CONVERT(NVARCHAR(2048), @ErrMsg)
    ;
 
or
 
@phrancis by the way, did you figure out what you're going to do with the query you want to execute against multiple server instances?
 
THROW 50001,'YOU SUCK!!!!!!!!'
 
@swasheck Nice!
 
@PaulWhite Coming as I do from a country where flags (and displays of same) can arouse (ahem...) considerable controversy, I do like the koru - is it relatively neutral? Personally, I'd prefer a different design for my own country's flag - rather than just the standard three blocks of colour.
 
5:37 PM
@Vérace Yes the Koru is pretty universal and uncontroversial. They'll always be someone, of course.
 
5:55 PM
@MaxVernon I have not yet, still working on the specific quirks first for data validation before I start deciding how to propagate this query onto multiple servers
 
0
Q: Should DBA be the primary phpMyAdmin usage support channel?

austinianThe phpMyAdmin support page points to stackoverflow.com and serverfault.com as the sole (non-bug related) support mechanism: Getting support Effective September 1st, 2013, support for questions directly related to phpMyAdmin has moved to stackoverflow. All prior support mechanisms have ...

 
@PaulWhite Would these "somebodys" be diehard "God, Queen, motherland (i..e. England) and only then country (i.e. NZ)"? The sort of person would regard the word "Aotearoa" as offensive?
 
Possibly. I've not experienced it personally.
 
@PaulWhite therefore it does not exist
 
0
Q: Editing a question to add a tag for RDBMS after the question has an answer

ErikIn the short time that I've been a member of the community I've noticed that the OP sometimes doesn't tag their question with the RDBMS (go figure :)). Community members often ask for this information in the comments and/or the OP leaks the data via their question or comments. I have edited a q...

 
6:30 PM
@swasheck It might not. It's just my experience that given any particular viewpoint, someone somewhere will hold it.
 
@PaulWhite Just like the web - if there's a perversion that you can even imagine in your most horrible nightmare, there's probably a website for devotees of the act in question. If you want a good laugh on the flag issue, try the "discussion" here.
 
The Guardian?! I hesitate to click it... ;)
There are some quite amusing comments there, thanks.
 
7:18 PM
Good question:
0
Q: Excluding a specific table from a SSDT deploy

Justin DearingI have an existing database with everything in schema dbo. I have an SSDT project with objects I am adding to it with schema foo I have a table that looks like this in the project: CREATE table foo.a ( id INT NOT NULL CONSTRAINT [PK_foo_a] PRIMARY KEY CLUSTERED CONSTRAINT [F...

I have no votes remaining today again :(
 
7:28 PM
Quick question on best practices - I'm refactoring some ad-hoc queries that have things like this...
select *
from supply_orders with(nolock)
Join CLIENT_EPISODES with(nolock) On epi_id = so_epiid
My initial thought was to just set isolation to read uncommitted (that's the same right?) since all tables are marked that way in the from clauses, and remove the (nolock) from the tables themselves
 
That makes sense, yes.
 
Yes if you really want read uncommitted behavior, it's the same. But you might need to take care to avoid this issue:
8
Q: "When a connection is closed and returned to the pool, the isolation level from the last SET TRANSACTION ISOLATION LEVEL statement is retained"?

FulproofThe MSDN online article "Snapshot Isolation in SQL Server" states: "An isolation level has connection-wide scope, and once set for a connection with the SET TRANSACTION ISOLATION LEVEL statement, it remains in effect until the connection is closed or another isolation level is set. When a conn...

 
I believe that is fixed now. Probably not in older versions. It's best practice to set isolation level explicitly at connection time anyway <-- all AFAIK
 
Yes, someone named twurtz who works for MS says it's fixed here: social.msdn.microsoft.com/Forums/sqlserver/en-US/…
In 2014, I mean.
 
Running 2008-R2, I'll try it and we shall see
 
7:40 PM
Oops, someone else from MS, not twurtz. But anyhow...
Erland pointed this out; there is a regression as of 2014 CU6.
 
Huh. Right it was 'fixed' in 2014 RTM then reverted to the original behaviour in 2014 CU6
Goodness knows. It's all very confusing. Glad I don't use it.
It took me a while to find the digit Erland left off the end of his KB link.
 
Me too. I was wondering what that other KB article had to do with SQL Server...
 
@JamesLupolt OT: our storage issue was related to a bug in Purity OS ... high latency on large sequential IO
 
@Phrancis you're aware of the pitfalls of read uncommitted and with (nolock) in general, right? dba.stackexchange.com/a/47364/10832
 
@swasheck Do you know what version of Purity OS you were running?
We found a nasty deadlock condition bug on large reads over iSCSI during our POC kit tests, but they fixed it last year and it was supposedly an iSCSI-only problem.
 
7:47 PM
@JamesLupolt Ha! Not much :)
I'd still rather set isolation level explicitly than use NOLOCK hints on every table reference.
 
Yes me too.
 
@JamesLupolt 4.0.20
 
actually @Phrancis - this is perhaps a better link: dba.stackexchange.com/questions/11571/…
 
RATTAIL JIMMIE HE'S A SECOND-HAND HOOD
 
I did work somewhere for a while where the tendency was to set the isolation level to read uncommitted and add NOLOCK hints everywhere.
 
7:49 PM
@JamesLupolt you can never be too careless
 
Unfortunately there is no resulting SUPERNOLOCK isolation level if you combine the two.
 
@PaulWhite Congratulations on getting the moderator position, may this be an opportunity for personal growth for you. I am sure you will do a great job. Cheers!
 
sycophant
(kidding)
 
@MDCCL Thank you :)
 
@swasheck I think we are still using 3.4. The sales engineer discouraged us from using 4.
 
7:53 PM
And I also want to thank the other candidates for nominating themselves, everyone has good qualities.
 
@JamesLupolt ours came with 4
 
@swasheck The nice thing is that it seems like they can fix bugs quickly, and install fixes with no disruption.
 
@swasheck ha ha I know you are kiddin'.
 
@JamesLupolt very true. it's just getting them to acknowledge the issue in the first place
 
@swasheck Yes, it took us a while to prove there was an issue. Once it got to the right people they were very efficient.
 
7:57 PM
my experience as well
 
I read something at the beginning of the year saying they only had about 1500 installations at the time. So we are probably all testing their stuff for them to some extent.
 
02:00 - 20:0020:00 - 23:00

« first day (1682 days earlier)      last day (3179 days later) »