« first day (443 days earlier)      last day (246 days later) » 

12:00 AM
That was a sleeper hit
It helped that everyone knew you
 
only half the room knew me
"This might be an unpopular opinion..."
gets five upvotes and 0 downvotes
ahem
@JoshDarnell any defense?
 
I'm sure everyone else knew you by reputation
Who has unpopular opinions? I'll report them to Twitter.
 
12:32 AM
Like READ UNCOMMITTED?
 
That's a very popular opinion with my clients
 
@JoeObbish the best I can do: I did say "might be" 😬
 
 
2 hours later…
2:28 AM
@ErikDarling Thanks for the dunk!
 
@Forrest it'll be one of the few i get on you
 
Also, bah, just found out neither of my sessions made the cut for SQL Saturday Dallas
 
you should have submitted some bullshit about ssis
 
SSIS... in the cloud
 
none of them deadlift
 
 
9 hours later…
11:22 AM
why doesn't the late answer review queue have the same comment options that the low quality posts one does?
it would be really helpful
 
11:55 AM
As I recall it's a question of triage and reputation level needed to access (and take actions on) the respective queues
If you find a low quality post during late answer review, flag it as LQ
 
hm
well, you know, when someone leaves a late answer that's just a blurb and a link, it'd be nice to be able to easily access the tsk tsk comment
 
yes I agree but that's not the purpose of the queue
I hear it is tough to keep people focussed
not really much of an issue on a site with small queue volumes but anyway
 
Canned comments for everyone! They taste better than SPAM! — animuson ♦ Sep 23 '12 at 16:19
 
but if it doesn't show up in the lqp queue...
meh, whatever
 
11:59 AM
@ErikDarling that's why I said you should flag it as LQ during late answer review
 
yes
suppose i'll just do that
 
I didn't write the system, I'm just doing my best to explain why I think it works the way it does
I might be wrong
 
i'm not arguing with you
 
I'm amazed it works, and you would be too if you saw the queries
 
you're right more than most people i know
@NickCraver if only you had access to people who were good at query tuning
 
12:01 PM
I hear Brent wrote all the queries himself
 
also created the ag
 
@ErikDarling /me hovers finger over banning button
 
BUT I'M SO CLOSE TO HAVING ACCESS TO SITE ANALYTICS
 
holds no fear when you consider who wrote the banning button SQL code
 
Badges are meh, but we decide on badges based on practicality of data access (e.g. sliding time ranges of "who was first or before X" is weighed...review queues, oof, they're just terrible - it's the only place that looks like business logic in SQL due to all the "can't review if...." constraints combined with what should be in there in the first place
 
12:03 PM
i hear the ban button user id value is hardcoded to 426
 
Oh and combined with an atomic queue of things people can't dupe review at the same time, so the list can't just be in SQL
 
you can still be tricky and open up lqp/late in new tabs and review the same answer twice
 
@ErikDarling wanna test it out and see? For science?
 
test it on lamak
he won't even notice
busy getting kissed by girls
 
@NickCraver It is amazing it works at all. I often think that when I see review items getting automatically invalidated based on some other action, and think how glad I am I don't have to maintain that code. SQL or otherwise.
@ErikDarling Did you just admit to abusing the system?
 
12:07 PM
i'm a white hat hacker, yo
i'm reporting a flaw
 
Has Sean fixed those AG issues yet?
 
oughtta be a badge for that
 
Any hacker worth their salt ought to be able to create one
 
i'll draw a picture of one
 
has worked before
 
12:09 PM
quite the doodler
hacking and doodling
what a life
 
renames user to Erik Doodler
 
the other day i typed my name wrong and it came out erik darkbug
> off by one
 
12:24 PM
@PaulWhite nope :(
 
Isn't that his whole job? What is he doing?
 
not deadlifting
@PaulWhite is there a technical reason for no parallel mods to table variables?
i mean they're all backed by regular old #temp objects
maybe they're just spools
🤔🤔🤔
 
12:47 PM
@ErikDarling might be as simple as no perceived use case
"table variables are for less than 100 rows"
 
yes but the query to get those 100 rows might be pretty rugged
 
also temp table is an easy workaround, though it doesn't work everywhere
yes, I understand that
that's why I said "perceived"
 
yeah joe tell me about easy temp table workarounds 🤭
 
have you run into a case where you couldn't work around it?
 
i was teaching this precon in madison...
 
12:50 PM
if an application doesn't allow temp tables it won't allow table variables either
 
blessing in disguise
 
I like table variables
any idea how few latches they require for creation? it's great
 
solved problem in 2019
 
we''ll see
 
teehee
how many latches does accelerated data recovery take?
gobs, probably
 
12:58 PM
@ErikDarling not that I can recall, no
 
well thanks for nothing
 
The dull answer would be that adding support for something requires work, which has to be funded.
 
just tell them oracle does it
 
There's no insurmountable difficulty I am aware of in making plans that write to table variables allow parallelism.
At least on the read cursor side.
 
yes
well
consultant bux anyway
 
1:04 PM
@ErikDarling tell them mysql does it ;)
Technically, it does. All mysql table variables use parallel plans.
 
man that's rough
 
 
1 hour later…
2:35 PM
Tell them Oracle does it and they'll do it
 
2 hours ago, by Erik Darling
just tell them oracle does it
 
You act like I read the transcripts
I'm too lazy for that
 
i mean it's like three lines up
 
Also, new Aussie slang learned from TV last night: F*** me dead = surprised
@ErikDarling 9 lines :)
 
not if you only count the user blocks
suckaaaaaaaaaaaa
go fix taryn's computer
 
2:38 PM
lmfao
 
is there a trace flag to change the hard coded 5 to a hard coded 10
 
@SeanGallardy I'm gonna be uploading another dump to our ticket for some of the ssms issues and can you give this person - twitter.com/Nick_Craver/status/1118988016889008144 access to the ticket?
 
2:55 PM
@Taryn Do you know their actual name, I'm assuming it isn't shuey bubbles and I don't have twitter :(
@ErikDarling Yes, it's 4
 
lol
i'm changing both of my kids names to shuey bubbles
 
@Taryn File is uploaded, I see it and am downloading it now
 
@ErikDarling you'd have to buy them shoes first
 
@SeanGallardy I assumed their name was shuey bubbles :shrug:
 
@PaulWhite this isn't nz we all have shoes
 
2:58 PM
@Taryn @SeanGallardy github.com/shueybubbles = David Shiflet
 
@PaulWhite +1
 
why'd he crop the bongos out of that picture
 
@PaulWhite Paul to the rescue :)
 
@ErikDarling is that an expression meaning excessive cropping?
 
no he looks like a BEATNIK
 
3:00 PM
Erik cropped the bongos out of that query
2
 
i'll have you lashed
 
to a bottle of wine?
 
WINELASH_OI
 
@Taryn He now has access and I'll ping him internally
 
@SeanGallardy gracias
 
3:02 PM
keep it family friendly, sean
 
@ErikDarling HEEEEYYYYYYOOOOO
 
blog post queue depth is now 2
 
on your top level or on sqlperf?
 
sqlperf is top level
 
i am aware
also they pay so
 
3:04 PM
do they
I should ask about that
 
i've heard rumors
i mean why else would rob farley write that eye roller abut filtered indexes
 
to delay my posts of course
typical aussie
 
shots fired
 
they'll land in desert
statistically
 
better than landing in the dessert
 
3:11 PM
true
newness++;
(Josh arrives)
 
why is everything interfering with me having my morning wine
 
poor planning?
 
incorrect
 
I've been summoned by increased newness.
2
 
I'm so new I didn't even get the newness smell on me yet, still waiting for it in the newness pipeline
 
3:20 PM
man
open source is rough
 
I thought open source was the answer to life, the universe, and everything?
 
Only if someone funds it
 
I have dictionary pressure for a CCI insert but the biggest dictionary is 1.2 MB
I don't understand
also I think it's the global dictionary
biggest non-global is 676 KB
 
sounds like a closed source problem
 
3:44 PM
Sounds like a dictionary problem, have you tried talking to Merriam or Webster?
 
top men are working on it
 
@Taryn The SSMS dump shows that we're waiting on a query for AG information to finished (hence the 10-15 minutes wait) but the message queue for the gui seems to be empty and running fine
I'll let David look into it more and update you all
It's also waiting on a WMI status check but I ignored that
 
@SeanGallardy ok, thanks
 
Interrogate Erik starts soon, someone ask his about distributed AGs :P
 
It would have done, but he's drinking and flying instead.
Alright folks, I got a flight tomorrow when I'd be doing Q&A. Your call.
 
3:54 PM
Erik is in a cab to the airport
 
Ah, tweeted the update that's why I didn't know
JFK?
 
Ouchies
Enjoy that line
 
70% voted for drinking
 
and $100 cab ride
 
3:54 PM
What line?
 
Also, even Josh is on Twitter, so it seems Sean's newness is finally confirmed.
 
checkin/boarding/security
 
$35 flat rate car service
Lol
I have precheck my line is 8 seconds long
 
Psssh I have precheck and the line is always 30 minutes long
 
Not my fault you don't look trustworthy
 
3:59 PM
I am shifty AF
 
can't believe Sean passed a background check
 
Only because of his scoliosis
 
@JoeObbish You have no idea how many I've been through
the government could probably re-create me from dna at this point
 
4:13 PM
@SeanGallardy I bet lots of people could
3
 
Is it me or has Joe gotten cheekier in the last few days?
It seems like a floodgate has opened
 
Seems he started drinking wine
 
@JoeObbish underrated chat message
Did Joe successfully drink wine?
I remember he mentioned it was in his 30 day plan.
 
4:29 PM
@SeanGallardy trying to plan our next steps - any word/movement?
 
4:41 PM
@ErikDarling it only seems that way because Sean disappeared for a week
 
Dunno about that
 
 
1 hour later…
5:46 PM
@NickCraver I have some ideas but I'm checking them at this point against a few things, I'm still waiting on the PG from their side though
 
ACK - any idea if that's likely today or days out? Most of our previous issues have been on-the-phone critical so I'm not sure what to expect here
 
@NickCraver I have it at a high priority right now, I'm on people to get back to me at mthe moment
 
:) <3
 
 
1 hour later…
7:20 PM
@NickCraver You should have some movement, check twitter please :)
We're working on it, I didn't forget you and @Taryn
 
looking but don't see it, DM?
 
I believe so
from Li
 
hmm see a ping from wednesday I need to follow-up on but nothing today
 
Hmm, okay. Well, know I'm working with them on it now and we're looking at stuff.
 
7:46 PM
"looking at stuff"
So technical.
 
8:07 PM
"trace flagsx
Sorry airport drunk
Actually cab drunk now
 
@SeanGallardy got a ping from Li saying something is in-progress - anything we can do to help? Unfortunately, all we're doing is a whoooooooole lot of waiting
 
8:20 PM
@NickCraver I understand Nick, right now, unfortunately no there isn't anything I need but I'm sure that'll change ;)
 
ACK
 
8:31 PM
@SeanGallardy FYI we're getting a series of minidumps it seems with:
2019-04-19 20:06:04.94 spid380s    Error: 41100, Severity: 16, State: 37.
2019-04-19 20:06:04.94 spid380s    The availability group '71333273-9f75-17f4-eadc-a6752dda7413' and/or its local availability replica does not exist.  Verify that the specified availability group name is correct, and that the local availability replica has joined the availability group, then retry the operation.
The last was 13 dumps in quick succession about 30 minutes ago
(may be unrelated, just a heads up)
 
8:49 PM
@NickCraver Would you upload one or two of those minidumps please?
 
absolutely - coming up
only 12MB zipped the whole batch + error log - uploaded for ya
 
Thank you!
 
9:05 PM
@NickCraver Nick, this is being caused by AutoSeeding it seems
Have you revoked the create database priv for the AG so it doesn't try to seed?
 
can we just alter the AG to turn autoseeding off, or need to go perms?
in a hangout and can call if easier - trying to parse now
 
We can try the ALTER but I'm unsure how long it'll take
@NickCraver whichever is easier for you
sqlmin!CHadrArBuildTargetStateCheckIfSeedingNeeded::SendBuildReplicaResponse
sqlmin!CHadrArBuildTargetStateCheckIfSeedingNeeded::SendSeedingResponse
sqlmin!CHadrArBuildStateBase::CallFuncAndResumeBuildTask<CHadrArBuildTargetOpContext,CHadrArBuildTargetStateCheckIfSeedingNeeded>
sqlmin!<lambda_141d5b3c3f92c35b3f9d77455d1eb80c>::<helper_func_cdecl>
sqldk!SOS_Task::Param::Execute
sqldk!SOS_Scheduler::RunTask
sqldk!SOS_Scheduler::ProcessTasks
sqldk!SchedulerManager::WorkerEntryPoint
sqldk!SystemThread::RunWorker
@NickCraver My phone didn't ring but I have a missed call from anonymous - Is that you?
 
@SeanGallardy aye
 
ok, let's try again lol
 
ok calling
answer your phone!
 
9:13 PM
ALTER AVAILABILITY GROUP [SENetwork_TAG] MODIFY REPLICA ON 'NY-SQL03' WITH (SEEDING_MODE = MANUAL)
[Verbose] 00000d04.00002f24::2019/04/19-20:54:14.528 INFO  [GUM] Node 1: executing request locally, gumId:238401, my action: /dm/update, # of updates: 1
[Verbose] 00000d04.00002f24::2019/04/19-20:54:14.529 DBG   [DM] Starting group transaction 25:25:25904678+1
[Verbose] 00000d04.00002f24::2019/04/19-20:54:14.529 DBG   [DM] Started single transaction 25:25:25904678+1::0+1
[Verbose] 00000d04.00002f24::2019/04/19-20:54:14.529 DBG   [DM] ClusDBop SetValue key: Resources\49652a7f-ae33-4877-b506-5345665ef665\Configuration, value: b56f322c-daa0-4558-b306-099f5ef93b23
 
176,048 seems like a lot of buffers.
 
You're a lot of buffers
 
I get that a lot.
Er, wait. That might have been a first for me.
 
9:32 PM
00000d04.00003548::2019/04/19-21:31:14.839 INFO  [GUM] Node 1: executing request locally, gumId:317043, my action: /dm/update, # of updates: 1
00000d04.00002d04::2019/04/19-21:31:14.840 WARN  [CLFS] Insufficient buffer. Reallocating and retrying (numRetries: 1, needed: 176048)
00000d04.00003548::2019/04/19-21:31:14.843 INFO  [GUM] Node 1: executing request locally, gumId:317044, my action: /dm/update, # of updates: 1
00000d04.00002d04::2019/04/19-21:31:14.844 WARN  [CLFS] Insufficient buffer. Reallocating and retrying (numRetries: 1, needed: 176048)
00000d04.00002d04::2019/04/19-21:31:14.835 ERR [CLFS] Log is full
 
10:28 PM
 
10:50 PM
@SeanGallardy we miss youuuuuuuuuu
 
@NickCraver @Taryn Soooo apparently my laptop has a thermal limit and I hit it...
 
haha
 
@NickCraver @Taryn
1. Remove the secondary AG from DAG
2. Remove all databases on the secondary AG
3. Add a secondary replica to the secondary AG
4. For any large size databases, manual seed them with log backup and restore on both replicas of the secondary AG
5. Join The secondary AG to the DAG
 
11:17 PM
I made it to the weekend
quite inconsiderate of me to say in here, I know
 
11:41 PM
Msg 41190, Level 16, State 7, Line 1
Availability group 'AG-SENetwork' failed to process add-database command.  The local availability replica is not in a state that could process the command.  Verify that the availability group is online and that the local availability replica is the primary replica, then retry the command.
SQL Server parse and compile time:
 
11:54 PM
but what was the parse and compile time?
 

« first day (443 days earlier)      last day (246 days later) »