« first day (442 days earlier)      last day (247 days later) » 

12:01 AM
YOU JUST WANT MY 109 SUBSCRIBERS
 
literally who.
once you take out people that subscribe to both of us
i wonder how many are left
 
I have only the most discerning subscribers
 
i'd be excited to pick up geoff patterson, maybe
 
the whole thing just seems scummy
 
I upvoted Joe's answer purely for the column name.
 
12:06 AM
"I see you've got all of these followers and blog posts. it would be a shame to lose them"
 
what about my answer
 
you bullied that poor azure user
B U L L I E D
 
tough love is necessary sometimes
 
@ErikDarling I didn't read it before because you implied it was a bad question.
Voted.
...to close the question.
 
good man
i can't believe i didn't put this in here yesterday while everyone was ignoring my urban playlist
 
 
1 hour later…
1:41 AM
I only ever listened to rap in the 90's. Is any of that stuff from the 90's?
Might have bled into the mid 2000's.
Forgot about Dre was cool.
2
 
2:35 AM
Forget Dre, this is the new cool:
Well, by "new", I mean 2 years ago.
Nice one, @Joe
 
2:51 AM
thanks
seemed like the kind of question you'd answer
 
 
1 hour later…
3:51 AM
Too long. A simple link to SQL Server Science would have been sufficient
 
 
7 hours later…
10:51 AM
which one of you starred the lox?
you're my new best friend
 
11:22 AM
Twas I
 
no one who stars the lox says "twas"
 
 
2 hours later…
1:03 PM
comment converter
You don't need a dynamic SQL query. Refactor as a UNION ALL with a different WHERE departmentkey=... in each SELECT. — Dan Guzman 3 hours ago
 
1:27 PM
I did what you said about the temp table using a memory optimized table. So, basically I inserted the records of an intermiedate result. That improved the query but then, when I had to join the memory optimized table to another tables, I had issues. I got to a point when joining to any table makes it bad. I cannot use table variable or temp tables because this query will finally be called from Tableau and that product has its limitations. — Francisco Goldenstein 13 mins ago
😬😬😬
 
@ErikDarling I took care of it.
 
@SeanGallardy ticket #119041821001133
 
double doubles
surely a blessed sign
 
@ErikDarling I see four dubs
 
yeah okay nice reading showoff
 
1:54 PM
That's, uh, some not-ideal throughput.
12k rows in 8 minutes.
 
You should see how long it can take to get less than 400 rows from an AG DMV...
 
I almost asked "is that in a DAG?" but thought it might be too soon =P
 
@JoshDarnell far too soon
 
wow
quest bought apex sql
a new redgate appears
 
@Taryn 😬
Sean will solve everything.
 
2:08 PM
I expect nothing less
 
@ErikDarling Sounds like there are exciting things going on with companies that make database-related products.
 
yes. yes indeed.
 
2:39 PM
@Taryn Got it
 
thanks!
@SeanGallardy dump file is still uploading
got about 3 minutes left
 
Aww I got your voicemail
:sad trombone noise:
Nice, got the file
 
@ErikDarling maybe in a few years they'll be one monitoring solution to rule them all
 
VOICEMAIL
that escalated quicky
 
who talks on the phone?
 
2:54 PM
Microsoft support specialists
they are v. friendly people
 
I'm downloading the file, someone didn't zip it first ;)
 
DIOS MIO
you can't ZIP things with sql server, sean
 
Right?
 
@SeanGallardy sorry it's been a week
 
Akshually
@Taryn I know you all have been up for 3 days
:(
 
3:05 PM
@SeanGallardy pretty much
 
It's still downloading... I'll let you know, would you rather me ping you here or we create a chat or I call/email?
 
need the file before chatting? Cause we can call ya in a few
 
@SeanGallardy pedal faster
 
I'm gonna zip it real fast and reupload
 
@Taryn Yeah let me look at it and we'll go from there
@Taryn No need :)
 
3:21 PM
I'm going to have to step away for a bit in about 10 minutes, but should be back in about an hour. If anything comes up before, feel free to ping Nick
@SeanGallardy zip file incoming
@SeanGallardy zipped file it up, I should be back in about an hour
 
3:39 PM
@JoeObbish i sure hope it's opserver!
 
OPserver? I don't like it already
 
😱😱😱
 
Opserver*
 
op's server?
 
Jan 11 at 3:27, by Joe Obbish
remember, OPs are always trying to trick you
 
3:45 PM
FWIW, we removed Opserver from monitoring the boxes at all and saw no change
 
i hope sean doesn't close your ticket as too localized
 
@NickCraver Background: there's a running joke in here that question askers (OPs) are always trying to trick us by providing misleading or incorrect information in their questions. The joke is carried further by assuming that anytime the substring "OP" is found it indicates some kind of subtle trick.
 
Oh and SSMS is absolutely terrible says sp_who*...we've found some super bad behaviors with all of the locking in there. For example, IntelliSense loops through every added database and queries all tables...regardless of if IntelliSense is disabled entirely...
@JoshDarnell =P
 
you should try azure data studio, i hear it fixes everything
 
So...I did
It does the same queries, except twice
 
3:49 PM
heh heh heh
 
it runs 2 in parallel, it's 2x as bad at the same problem
 
πŸ‘€
 
But, not having the Synchronized, etc. indicators, it didn't contribute to the locking as bad as SSMS
 
@JoeObbish prolly yes, but you're not alone.. just spilled some beer over the screen :D
 
hooking SSMS to an empty server and profiling is...kinda scary
 
3:50 PM
yeah, the ag dashboard query in ssms is the hottest of messes
 
oh it's not the dashboard, just the DB list background stuff
 
I hear Aaron is only using Azure Data Studio from here on out.
 
i'd imagine sqlprompt and other tools have the same problem
activity monitor also runs bad things
it's sort of amazing how bad microsoft is at sql server
3
 
There's a reason I spent a lot more time than I ever normally would in SQLCMD yesterday :(
 
you have to see idera's queries..
 
3:52 PM
nope
 
they even have the option to hide their own shit
 
would rather learn foxpro
 
if you have RG enabled there might be a way to black hole those queries
I bet nothing would go wrong
 
block by app name? :)
@JoeObbish sorry if you're not used to have dumb questions from your subscribers.. but related to this question:
12
A: If a VARCHAR(MAX) column is included in an index, is the entire value always stored in the index page(s)?

Joe ObbishValues that exceed 8000 bytes cannot be stored "inline". They are stored on LOB pages. You can see this with sys.dm_db_index_physical_stats. Start with a simple table: DROP TABLE IF EXISTS #LOB_FOR_ME; CREATE TABLE #LOB_FOR_ME ( ID BIGINT, MAX_VERNON_WAS_HERE VARCHAR(MAX) ); CREATE INDEX IX O...

running your code fails for me in SQL 2016 SP2
sys.dm_db_index_physical_stats(DB_ID(), OBJECT_ID('#LOB_FOR_ME'), 2, NULL , 'DETAILED');
Msg 2561, Level 16, State 11, Line 14
Parameter 3 is incorrect for this statement.
I'm most likely missing something.. but don't see what
(that 2 is not an accepted value as far as I know and as per the docs..)
 
4:07 PM
docs say it's ok
index_id | 0 | NULL | -1 | DEFAULT
Is the ID of the index. index_id is int. Valid inputs are the ID number of an index, 0 if object_id is a heap, NULL, -1, or DEFAULT. The default is -1. NULL, -1, and DEFAULT are equivalent values in this context.
"Valid inputs are the ID number of an index"
did you create the NCI on the table? if so, what's the index id of it?
 
@Marian It needs to be executed in the context of tempdb.
Since Joe used DB_ID() as the first parameter.
 
@JoeObbish yep, created it, ran the entire code
@JoshDarnell aaaa that might be it.. was running over master
 
πŸ‘
 
I changed the answer to make it @Lamak-proof
3
 
Haven't heard from Lamak in a while.
 
4:11 PM
@JoeObbish it was 2, no need to call @Lamak .. I can be dumber without effort
 
I didn't mean it in a bad way at all. He's a very discerning guy
v. detail-oriented
 
persnickety
 
@JoshDarnell he's here, I can sense his kissed chickbone running a background thread
(he had an avatar in the past with a red kiss on his cheek)
@JoshDarnell yeah, apologies.. told you it was a dumb q
 
@Marian Hey, no problem! I was puzzled by it as well until I started messing around with it in SSMS after I saw your question.
@Marian I appreciate you consideration for my newness (I didn't know about that old profile picture).
 
I didn't know that either
bring back classic lamak
 
he was 4 years old in that picture
 
he still kept it, probably a 10 yo pic :)
 
0
Q: How can you have a Buffer cache hit ratio of 100% with PLE of 103?

James JenkinsI found What does a "Buffer cache hit ratio" of 9990 mean? (I was getting result of 9000+ from a blog that was wrong) and used the solution by LowlyDBA/Denis Gobo combined with the PLE checker from Denis Gobo and got a Buffer cache hit ratio of 100% with PLE of 103, how is that possible? The b...

> My instance is SQL 2017 CU13, heavy OLTP, with 4GB of RAM for the SQL instance (single instance, 6GB RAM on the server).
"heavy oltp"
sean could deadlift that
 
@ErikDarling working with data takes its toll
@ErikDarling didn't know sql can run on a phone
 
4:44 PM
@Forrest
> Buffer cache hit ratio and Buffer cache hit ratio are measurements for a very small timeslice.
i think you think you might have repeated yourself junior
 
err, thanks
fixed
 
@ErikDarling 15 lbs or less!
 
@Marian what a stud
 
i'm surprised you can lift a leg to pee
 
Sean has competition
well, he used to
 
4:49 PM
Only for you
 
Are Lamak's wild days behind him?
 
wife, kid, condo
 
sounds like it
 
"PLE: 45,876" - I think there is enough RAM on this server.
 
@JoshDarnell Never
 
4:52 PM
that's like saying there's enough wine in the bottle
2
 
or that Joe has written enough blog posts
 
wait
has he?
 
5:07 PM
@SeanGallardy around for us to chat with ya?
 
Breaking: Top Microsoft Man MIA for RCA
 
@JoshDarnell he's OOF
 
5:30 PM
is that a sith?
 
NAh we had a old technology phone call
"Who uses phones anymore?"
My rotary phone doesn't get twitter
 
joe still uses smoke signals
 
5:47 PM
@Taryn You're going to SQL 17 CU 14 correct?
 
@SeanGallardy yes. And the impacted server is already CU14 - the other is CU 12
 
6:08 PM
smart to skip CU 13
everyone knows that 13 is unlucky
 
i have yet another exchanges spill demo that works every ~10th time
very unprofessional
 
every 10th time? interesting
 
well
about every 10th time
hence the squiggle
 
6:30 PM
Very unprofessional indeed.
That would be pretty funny in a presentation though.
"Okay, so here's an Exchange spill. Er wait, let me try that again. Okay, one more time..."
 
@JoshDarnell It seems I've been replaced by jcolebrand and Josh!
(cole handled the comment flag)
 
did you clap?
also I just spotted a parallel merge join
on my system?!?
I'm going to eradicate it
 
6:47 PM
wew lad
 
7:00 PM
yeah
MAXDOP 1 is 4X faster
even if I keep the merge join
Thanks @SeanGallardy
 
@JoeObbish in my mind a little maybe
 
@JoeObbish I do what I can :)
 
7:53 PM
@NickCraver @Taryn Alright, everything is submitted and such, waiting to hear back
 
@SeanGallardy thanks!
 
submitted?
i thought you were the escalation engineer
 
he submitted to the robot that tells him the answer
 
I'm just a pleb
;)
 
Is it ML?
 
7:56 PM
Everything is apparently ML these days, it's the cool buzzword
 
MalAISE
 
becky disagrees
 
@ErikDarling ...literally who?
 
0% hoodrat
 
Josh is a good southern boy
 
8:37 PM
@JoeObbish I think that's the first music link I've ever clicked on in here.
I'm not sure how I feel about it.
 
suckerrrrrrrr
 
Joe is a Trusted User.
I thought I was safe!
 
Paul is an extra safe user. You can always trust his links.
 
Notable dunkage in progress:
5
Q: Why doesn't SQL Optimizer use my constraint?

ZikatoI want to create a table with a NOT NULL bool column. I use TINYINT with CHECK constraint BETWEEN 0 and 1. The constraint is new and thus trusted Now I would expect that SQL optimizer now knows that this column can only be 0 and 1 so when I write query col >= 2 I will see Constant Scan in actua...

@Forrest -__________-
 
8:53 PM
brent was sort of right
 
in The Heapβ„’ – Consultancy ©®, Jan 23 at 13:15, by Paul White
be sure to check out the opinions on http://meta.stackoverfiow.com/questions/260245/code-edits-revisited
 
@ErikDarling How so?
 
auto parametrization is only considered for trivial plans
 
This is what dunking is all about.
I've never heard of CREATE RULE and sp_bindrule before.
 
Welcome to SQL 2000
 
9:10 PM
@JoshDarnell you have good taste
@SeanGallardy top microsoft man recommends SQL Server 2000
 
@JoeObbish To some people, that's still an upgrade :X
 
> mfw he does it again
 
@SeanGallardy if that's your fix, you're fired
 
FWIW, I heard it doesn't block on AGs though...
 
9:18 PM
@SeanGallardy MySQL users are not people
 
@ErikDarling What's the query that does that parallel merge join you Tweeted about?
 
> not being able to reverse engineer it from the query plan
 
The newness has not yet worn off.
 
9:39 PM
I've got a punching bag with a picture of a parallel merge join plan on it
 
@Taryn Every day that I'm not fired, I'm surprised ;)
 
sounds like my job
 
10:06 PM
anyone know if the formula for the auto stats percentage was worked out by anyone? if so, what is it?
or at least what the inputs are? I assume row count and page count?
 
new or old?
 
not auto update
as in if I update stats and I get a 0.2% sample
what's the formula for 0.2%?
I suspect it last changed in sql server 2008, don't know for sure
 
oh
didn't paul have a q&a about that?
 
26
A: Strange behaviour with sample sizes for statistics updates

Paul WhiteBackground Data for the statistics object are gathered using a statement of the form: SELECT StatMan([SC0], [SC1], [SB0000]) FROM ( SELECT TOP 100 PERCENT [SC0], [SC1], STEP_DIRECTION([SC0]) OVER (ORDER BY NULL) AS [SB0000] FROM ( SELECT [TextVa...

thinking of that? it doesn't answer my question
 
that was the one
probably
you should leave a comment for paul about it not answering your question
 
10:41 PM
minidunk
0
Q: Converted a Scalar function to a TVF function for parallel execution-Still running in Serial mode

user9516827One of my query in was running in serial execution mode after a release and i noticed that two new functions were used in a view which is referenced in the LINQ to SQL Query generated from the application. So i converted those SCALAR functions to TVF functions,but still the query is running in se...

 
I'll just figure it out
I just need approx anyway so it should be ez
 
11:05 PM
This is where Joe gets dunked on by the update stats algorithm.
 
I've heard Joe can apply math.
 
I'll do my best
also, I hate that you pointed out that perf issue with creating an index
I SEE IT ALL THE TIME NOW
 
@JoeObbish 42
 
11:40 PM
Lol
 
11:53 PM
I need a regression solver
 
Send a repro to Sean
He'll submit it
 
my I/O can't keep up :(
 
With what?
 
the animu girls of CDM would be ashamed
I'm gathering data to figure out the stats issue
 

« first day (442 days earlier)      last day (247 days later) »