« first day (57 days earlier)      last day (632 days later) » 

12:01 AM
get what I mean now?
 
wtf did I just watch
also that game needs better hit detection
 
sean-chan
when you make comments like that...
anyway
 
I bought a somewhat similar game for the xbox 360
for like $5
it arrives in the mail
there's this extra stuff
 
@JoeObbish What you do with the human sized pillow is your choice
 
12:05 AM
there was a pillowcase
with, uh, anime girls on it
I really had no interest in such a thing
I threw it on the ground to deal with later
 
LIES!
You put it on the pillow and cuddled with it all day
 
it's not like I was going to have company over any time soon, so why did it matter?
the next day I'm waiting at the bus stop to go to work
bus-riding friend is there too
it's winter and the weather is pretty bad
bus never shows up
I suggest that we wait in my apartment for the next bus to show up
since it's just a minute walk away
 
Did you rufi them?
 
and it's warm
We enter... and he SEES IT
he sees the anime girl pillow
I never spoke to him again
 
Your life is over
 
12:08 AM
"I... I was going to th-throw it away, I swear"
 
"Is that a semen stain?!"
 
keep it family friendly
 
love stain, please
 
Erik must "love" many things
 
i am incapable of such frivolity
 
12:12 AM
anyway
erik, you believe me, right?
 
@sp_BlitzErik He says, after calling it a love stain
 
@JoeObbish i can't believe you made a bus friend
 
that's not that bad, right?
 
None of the women have three breasts or a penis so you're safe
Looks like a nice silk pillowcase to keep you cool on those long sweaty nights
 
@sp_BlitzErik why's that? it's not like NYC where everyone wishes you were dead
@SeanGallardy-Microsoft this guy expects anime pillowcases included in $5 games to be made of silk
come on now
I actually own silk pillowcases
not $5
 
12:17 AM
silk is awful to sleep on
 
Hey if it's real silk, I'd gladly pay $5
 
@sp_BlitzErik wut
 
For sheets? Yes. Pillowcase? No.
 
what's wrong with silk sheets
 
Sheets you end up sliding all over and then eventually sinking into your own pool of sweat
 
12:18 AM
I don't have that problem
 
so... many... comments...
 
I sleep like a king
 
too soft, too hot, and very unruly
 
I sleep in a king... sized bed
 
it is a bit unruly
well what do you recommend?
 
12:20 AM
i have very particular tastes that do not often suit people
 
does the wife put up with it?
 
i like rough, stiff sheets
 
rough... sheets...
like sandpaper?
 
basically
threadcount: 2
 
have you tried sleeping on a rock?
maybe you were born in the wrong era
I bet you'd make a good caveman
 
12:22 AM
well i'm not a terribly good modern man so
who knows
 
@sp_BlitzErik I've heard that about you
 
@SeanGallardy-Microsoft what kind of sheets does the lady killer have?
 
"erik, bungabase slow. glug no count meat while grog eat meat. bungalock."
 
When I lived up north, I liked lower thread count sheets for the summer and higher thread count for the winter. Summer was 600 TC Pima cotton or similar, sometimes going to a blend. Winter was 1200 or higher TC Long Staple Egyptian Cotton.
 
I don't think I've ever met anyone as fancy as you in real life
 
12:25 AM
LMFAO
Don't mess with my sleep
4 things motivate me, in total, in life
sleep is one of them
 
sleep, women, tea
hmmm
I'm short one
weightlifting?
 
you have 2 of 4, you know sleep is one
 
steak, gym, women, sleep
 
steak seems too recent
sql server?
 
12:31 AM
I'd thought you would know better, food doesn't motivate me - I can't eat any of it. I eat to survive, not because I like it.
 
the guy took the MCM for fun
it's not that unreasonable
 
LOL
No SQL Server isn't one of them
 
travel?
paying as little taxes as possible
 
Maybe some day, over a nice steak dinner, I'll tell you the other 3
 
hey I got 2
 
12:33 AM
You're unsure what the second correct reason is, though
 
it's tea
 
anyway, "motivate me" is not clear in the first place
motivate you to do what? live?
 
To do anything I do in life
 
it's obviously not "what gets me out of bed in the morning"
well FREEDOM must be one of them then
 
12:34 AM
Not directly though it could impact some
 
> be american
> claim that FREEDOM doesn't motivate you and isn't one of the most important things in your life
> mfw
 
"Freedom" is subjective
 
only for commies
anyway it's your life, nothing wrong with being subjective
 
Freedom of speech? of opinion? of press? of????
Freedom to own my one (1) gun? (for the cia that's watching this chat, I only own 1 (one) gun)
Did you email Erick back?
 
12:55 AM
sure, all of the freedoms
 
that name is confusing me
can you call him something else
 
it's a fake name
pretty sure
he chose it
anyway, that blog post is looking less and less likely
 
quit goofing off in here
sean is out looking for his wheyfu
 
I'm still sick
sick people can't blog
doctor's orders
 
sick blogs are the best
 
1:07 AM
I'm seeing a query plan with a DISABLE_OPTIMIZER_ROWGOAL use hint
 
but the cost of a scan is still reduced by the row goal
I thought that shouldn't be possible?
 
Couldn't find my iron... turns out I must have lost it in the move 1.5 years ago.. so it's been 1.5 years since I've ironed anything.... I'm winning at life for sure!
 
did tf 2363 get nerfed?
 
@sp_BlitzErik what makes you say that?
@SeanGallardy-Microsoft you're smalltime
I've never ironed anything in my life
I've also never owned an iron
 
1:11 AM
CA had dry cleaning across the street from me, so it's probably been more like 6 years
 
I'm pretty sure if I tried I'd light something on fire
 
@JoeObbish bows
 
I owned a 9-iron once
does that count?
 
close, but nope
 
i'm an idiot
3
forgot to switch my db to 2017 compat
 
1:13 AM
rookie
 
@SeanGallardy-Microsoft make a restore switch to upgrade to the latest compat level
thanks
@PaulWhite someone got my laptop with all the right settings
 
@sp_BlitzErik If I were the compatibility CZAR I totally would
 
i don't wanna hear it
 
Good thing you're reading it, then
 
compatibility mode is weird
I've never understood it
are there any MCMs around to explain it?
@sp_BlitzErik it's unanimous
 
1:15 AM
@JoeObbish Looks at Paul White
 
I thought Paul was one past an MCM
MCMCM?
 
I have no qualifications
 
not true
you are the man on the $100 bill
 
This medical form asks if I've ever broken a bone and give a 2cm space to write down which... amateurs
 
"most of them"
 
1:17 AM
Speaking of which, I have a bunch of query bucks - can I buy a voltron ring or something?
@JoeObbish good one!
 
ok let's get serious
anyone else read CU fixes for fun? is that just me?
 
I read them all yeah
 
I read them all
 
off the top of your head, do you remember anything post 2016 SP1 about DISABLE_OPTIMIZER_ROWGOAL?
also this is definitely the best chat room
 
@JoeObbish Serous is it only gives me 7 lines to write hospitalizations and surgeries
 
1:19 AM
3/4 people read CUs for fun
 
Like what if I have around 22?
 
@SeanGallardy-Microsoft how far back does it want you to go?
 
"You've had"
to me, going through the Paul White optimizer, means forever since this exact moment in time
 
can't you just refuse to answer?
you're an American
 
Yeah, but I'd like my level of care to be above Canada's
"Concussions" plural
 
1:22 AM
@JoeObbish Sorry I was eating a sandwich. No I don't recall anything.
 
a test for the shop?
 
I'm also answering in the most "obtuse" way
"Do you drink?" -> Yes. "Frequency" -> biannually
 
just refuse to disclose it for religious reasons
 
No, I want to see these people struggle with real words
 
@JoeObbish No just a regular lunchtime snack
 
1:25 AM
in case anyone is willing to take a look: brentozar.com/pastetheplan/?id=ryZwl6Kqz
here's the query hint: OPTION( USE HINT ('DISABLE_OPTIMIZER_ROWGOAL'), HASH JOIN, LOOP JOIN )
I'm looking for a yes/no mostly
but that query plan doesn't seem like it should be possible
look at the operator cost for node 5...
 
@JoeObbish To what question?
 
is that plan reasonable given the use hint?
 
What is the top expression? Top (1)?
Jeez I hate anonymized plans
 
TOP 1
sorry, I know
no choice
 
Well then: yes.
 
1:31 AM
night all - don't do anything I wouldn't do
 
@SeanGallardy-Microsoft I just ate crackers
 
I'm not even fully gone and you already drop the ball
smh
 
Yeah hey that's pretty restrictive
 
THIS IS WHY WE CANT HAVE NICE THINGS
 
@PaulWhite My thinking was that the row goal hint should prevent the row goal related discount from getting applied to the clustered index scan at node 5
 
1:32 AM
Let me rephrase, "Don't do anything, outside of food, that I wouldn't do"
 
Well it's why you can't have nice things
 
hehe, tip-o-the-hat
 
and the subtree cost should be IO + CPU as usual instead of 0.0127238
what am I getting wrong?
 
@JoeObbish Ok fine but you still only want one row. The cost of the top is going to cap the subtree cost, given the lack of a blocking operator.
How about putting a repro together.
 
damn
there's so much that I don't know
and you're right, to be fair I did just ask for a yes/no
 
1:35 AM
Well I'm not going to say it's always defensible or sound logically.
And it's not always straightforward either.
There are some interesting cases around operator cost, exploration, pruning and whatnot.
 
well, it's not so bad
in that I would expect an OPTIMIZE FOR or maybe even a FAST to solve the problem
at least there are immediately obvious things to try...
 
With a top (@var) sure
 
I guess I thought that TF 4138 (and equivalents) did more than they actually do
 
So why the question about the hint in CUs?
Just out of interest
 
I remember a sequence of events
where we had performance issues caused by the TOP
as in, we got the nested loop plan
not a good choice here
we added the hint and the plan changed
the customer seeing this is on 2016 SP1 base
 
1:41 AM
Is that still supported?
 
I wish it wasn't
 
Sorry yeah ok I see
 
I even wrote a powerpoint presentation about the evils of 2016 SP1 base
 
Of course you did
 
I did, I wrote it today even
it's very short and I'd be happy to send it to you
 
1:41 AM
Yep
 
I'm not joking!
 
I know
I meant "of course you did" because you've done nothing but PPs recently
 
anyone using columnstore, in-memory, and probably a few other things really needs to stay on top of CUs
it's rough out there
 
Yes things have definitely changed
Could be worse
Like finding a stable SSMS release
At least box product CUs tend to be stable and usable, if not optimal
 
what are you on?
 
1:43 AM
17.6 now
 
oh, were you on 17.5 recently?
 
Yes I've been good about upgrading SSMS
And we had early access as well so
 
I heard bad things from coworkers
I've been too lazy/busy to upgrade
 
17.6 is the best of the 17.x line I would say
Unless you're into maintenance plans
I suppose I should install SOS to reset my expectations
 
SOS?
hmm. I don't even have sql server 2016 on this machine...
 
1:45 AM
SQL Operations Studio
 
Paul installs 2008 R2 - 2017 and I only manage 2017
shameful
 
tbf I use 2017 almost exclusively
The other instances only have AdventureWorks right now
There was a time when I would try to remember what was introduced when
Those days are pretty much behind us
Such is the rate of change
 
I actually don't know how to repro this one
I'll try it as an EXISTS query first because that's so much easier
 
And you're absolutely right about keeping up to date if using recent tech
Must be a nightmare if you're on the bleeding edge with high spec hardware
And dodgy queries :)
 
I'm finally right about something!
things are dodgy even on the latest CU
 
1:49 AM
Of course. New normal.
 
I have a support ticket open that I cannot say anything about publicly
 
Since when are support tickets under NDA
 
they can be
 
Oh because your company wants it secret?
@JoeObbish Have you signed something?
 
more like, I think that a proper professional would not comment on it
 
1:50 AM
Is he in today?
 
I can... kind of... imagine myself in his shoes
 
TF 1239
we ran into a production issue for which that one helped
support gave it to us
we had to ask to share it with our customers
they took a few days and agreed that was ok to do
 
But right, we can't talk about it
 
I never got that context actually
I was rather surprised that there was a trace flag for it
I assumed it was related to some other fix and we just got lucky
 
1:53 AM
Most things that might need to be switched off in an emergency get covered by a TF
 
Or a feature switch
 
that was the proper resolution for our issue
@PaulWhite that makes sense. so it seems as if they did a fix that they knew was risky
and this was an unintended consequence
 
Well it's only sensible to have a kill switch
 
that was the day that I soured on MVPs as a general concept
 
1:54 AM
Everything's risky really
 
well, that's not true. with MSFT's implementation of the MVP program
 
@JoeObbish Why so?
 
Our customers ran into an issue that was difficult to troubleshoot. I spent a fair amount of time coming up with a repro. We did the right thing and opened a ticket with support and got an emergency TF to help. we recommended it to our customers on the affected versions
One of them asked an MVP about it (who has access to more information about it than us), and the response was, "No, don't use if it isn't documented"
How is that being helpful?????
 
Well tbf he/she might not have known about it
And MVP says nothing about technical competence
 
I heard that it was mentioned in the secret notes of a KB and there was an email discussion about it on the MVP list
that might not be true though
@PaulWhite Yes, but many of the perks of the MVP program seem to imply that
 
1:57 AM
It is true, but not everyone partakes or reads every mail. There are a lot.
@JoeObbish Do they? People have been saying MVP != technical for as many years as I can recall.
There's a lot of variance for sure
 
@PaulWhite right, and I don't expect them to
but then don't make a comment like that... it just isn't helpful
 
And no one really knows the award parameters anyway
 
it can be hard for us to overcome that kind of thing
@PaulWhite People outside of MSFT say that, you mean?
 
@JoeObbish Yes, MVPs say that.
 
yeah, that's fine
but as far as I can tell, the perks of the program don't reflect that
for example, the additional information in the KB articles
that's a very technical thing, right? (ok if you can't comment)
 

« first day (57 days earlier)      last day (632 days later) »