The Heap™ – Consultancy ©®

Maybe one day these systems will understand code a little better and be like, oh you wrote a new function and someone else wrote a new function in the same place. They are both new functions, I should just keep them both since they were originally from two different branches.

instead of telling you there's a merge conflict

Paul White

git must be great because Microsoft paid billions for it and we know they only ever make great investments

Sean Gallardy - Mostly Retired

Like: Yammer?

Blizzard Activision

Nokia

Paul White

1:48 PM

Yeah

Sean Gallardy - Mostly Retired

Jinx, you owe me a beverage

Paul White

Sean Gallardy - Mostly Retired

:shudders:

Paul White

@ErikDarling or jerk you around

dezso

@SeanGallardy-MostlyRetired well, without a process it can be bad indeed. Fortunately there are solutions (workflows) to that.

Sean Gallardy - Mostly Retired

1:50 PM

I'm literally resolving merge conflicts such as:

>>>>>> HEAD

========

<<<<<<<

wow, glad we resolved that whole whitespace issue

dezso

on the other hand, most errors are solvable. I still remember tree conflicts from SVC which needed to copy the files elsewhere, purge the whole repo locally, then do some magic to restore order

Sean Gallardy - Mostly Retired

😬

Sounds like a Microsoft product

dezso

@SeanGallardy-MostlyRetired I feel your pain. This is not really git's problem, however.

Erik Darling

@dezso i still do that with git sometimes

dezso

@ErikDarling there are times when it's easier that way, I admit

Erik Darling

1:52 PM

yeah

Sean Gallardy - Mostly Retired

@dezso I just believe (which I realize doesn't mean anything) that it should be smart enough to not waste time on this stuff.

Erik Darling

i can't be bothered to spend an hour figuring out git. i use it to make open source stuff easy. as soon as it's hard, i'm bailing. it's not my full time job.

dezso

@SeanGallardy-MostlyRetired you could probably have a hook that consolidates whitespace-only lines

it probably helped me a bit having colleagues earlier who really knew git

Erik Darling

you may have responded to the wrong person, which is causing a merge conflict

dezso

oops

it didn't take long I got busted, sorry @SeanGallardy-MostlyRetired, now you know whom to blame

Sean Gallardy - Mostly Retired

1:57 PM

😂🤣

Paul White

2:09 PM

who to blame, or on whom to pin the blame

anyway

is there anything more frustrating that dealing with people who think they know heaps more than they actually do

there are plenty of things I don't know much about, but at least I am aware of that reality

Hannah Vernon

@ErikDarling amen lol

This is a new one for me:

ypercubeᵀᴹ

@SeanGallardy-MostlyRetired merges are often hard in git, indeed.

Hannah Vernon

@Charlieface the server is 17763.3287 (Windows Server 2019 Standard, Version 1809)

ypercubeᵀᴹ

The more the two branches have diverted, the more hard it gets. There are a few merging strategies (that you could experiment with). Also diff algorithms (which are used by merge) and options to ignore various white space changes. git-scm.com/docs/merge-strategies

Erik Darling

2:29 PM

@PaulWhite link

Sean Gallardy - Mostly Retired

@HannahVernon Running out of memory, eh?

@ypercubeᵀᴹ It seems my life has quickly become, "Do I merge or rebase?"

Paul White

Aug 30 at 10:08, by Erik Darling

@PaulWhite all of their problems are self-inflicted and far too localized and you can never talk them out of doing the thing that’s causing a problem.

Hannah Vernon

@SeanGallardy-MostlyRetired 2.8 GB available out of 8GB on this particular VM on my workstation.

Erik Darling

@PaulWhite stop talking to my clients

Sean Gallardy - Mostly Retired

@ypercubeᵀᴹ Sadly there are hundreds of people all working on stuff and I don't get to choose the options but it's good to know they are there!

Paul White

2:34 PM

@ErikDarling ah well if you're on dba.stackexchange.com/users/134691/searle1986's case I'll leave it to you to invoice

Sean Gallardy - Mostly Retired

@HannahVernon The only way that the stack can't make a new guard page (grow) is a stack overflow is hit or there isn't enough memory

ypercubeᵀᴹ

@SeanGallardy-MostlyRetired I surely hope you don't have to merge or rebase 100 different branches !

Erik Darling

@PaulWhite yikes no

Sean Gallardy - Mostly Retired

@ypercubeᵀᴹ My branch(es) from master, which has many PRs each day... meaning I'm almost always going to hit a conflict.

Hannah Vernon

@SeanGallardy-MostlyRetired it's running a C# app that I created, so I wouldn't be surprised if there is some bug somewhere. Also, I blame Hyper-V lol (I don't really but I have to blame something)

Sean Gallardy - Mostly Retired

2:37 PM

Is hyper-v using dynamic memory?

Hannah Vernon

it is, so that's probably it. However, as I said the host is not low on memory.

I mean it's nothing critical, I just thought it was an interesting error message

Paul White

"Unable to grow the stack to due memory pressure" would be a more helpful error message

Hannah Vernon

On another note (and indeed a physically separate host and VM), any reasons off the top of your head why SQL Server 2019 CU17 cannot ever be shutdown via the service manager or SQL Server Configuration Manager? I am forced to log into it, and issue a shutdown with nowait; manually to get the service to shut down. I made the mistake of attempting to restart the machine without doing shutdown with nowait; a couple of days ago and had to kill the power to get the machine to shut down.

this is a new install on a new Windows Server 2019 install, and nothing except me is connected.

personally I have never ran into that before

Sean Gallardy - Mostly Retired

shutdown causes all databases to checkpoint, my guess is checkpointing the databases is taking a long time, hence with nowait works.

Hannah Vernon

I was under the impression that the engine issued a shutdown with nowait when it receives a shutdown message from the Service Manager

Sean Gallardy - Mostly Retired

2:43 PM

no it's a regular shutdown

Hannah Vernon

are you sure though :-O

Sean Gallardy - Mostly Retired

very

Paul White

@HannahVernon docs.microsoft.com/en-us/sql/t-sql/language-elements/….

Hannah Vernon

so I should see some system process in sys.dm_exec_requests doing a checkpoint after I do a service shutdown then? And how long should checkpointing last if there is literally nothing going on. I'm going to test it obviously.

> Immediately stops SQL Server.

very clever

Paul White

@HannahVernon The link should have highlighted some text 4 u

Hannah Vernon

2:47 PM

Firefox doesn't use Chrome

Paul White

> Other tools and methods can also be used to stop SQL Server. Each of these issues a checkpoint in all databases. You can flush committed data from the data cache and stop the server:

By using SQL Server Configuration Manager.

By running net stop mssqlserver from a command prompt for a default instance, or by running net stop mssql$instancename from a command prompt for a named instance.

By using Services in Control Panel.

Hannah Vernon

but yah I did read the entire page

and it's not exactly my first time shutting down a SQL Server instance

Paul White

well it literally says those methods issue a checkpoint in all databases, and you seemed unclear on that point

6 mins ago, by Hannah Vernon

I was under the impression that the engine issued a shutdown with nowait when it receives a shutdown message from the Service Manager

seems like something someone running shutdown for the very first time might say

Hannah Vernon

hah

Paul White

but anyway, if it were me, I'd try running a manual checkpoint in all databases

they'd normally be configured for indirect checkpoints though?

something unusual is going on, but it's hard to guess what, beyond what Sean said

ypercubeᵀᴹ

2:51 PM

@SeanGallardy-MostlyRetired the options I mentioned are available when you do a merge/rebase in your local repo. Unless you have to use some kind of centralized intreface

Hannah Vernon

let me test some thing

Paul White

It somewhat reminds me of people complaining about status icons not appearing for servers in SSMS. It's always possible to fix, but no fun at all.

Sean Gallardy - Mostly Retired

@PaulWhite I really dislike this feature, tbh.

Erik Darling

i was thinking recently that it would be nice if you could do some filtering about which plans to collect in ssms when you get actual execution plans

so like if you're running a big stored proc you don't crash ssms getting everything

Paul White

@SeanGallardy-MostlyRetired Yeah it's flaky as all hell

Sean Gallardy - Mostly Retired

2:55 PM

It's always a windows permissions issue or a wmi issue (whether those are caused by other items like firewalls, etc., doesn't matter)

Paul White

Yeah no fun at all

Especially WMI

@ErikDarling SSMS is stupid for trying to render so many plans in the first place. As a 32-bit executable, it ought to be more aware of its own limitations

That said, recent SSMS builds have started running out of window handles every few days for me regardless, so I just restart the damned thing every so often.

Red Gate Tab History does a good job of recovering the tabs I had open, except the pinned status

Still better than ADS

The number of times I have forgotten to turn off plans before running a cursor 🤨

Erik Darling

heh yeah

Charlieface

3:11 PM

@HannahVernon Is that the client machine, the SQL Server machine or the DC? Or all of them?

Erik Darling

3:36 PM

"brent said i should use ctes"

Sean Gallardy - Mostly Retired

negative 50000 zoom

Paul White

4:00 PM

crikey

Hannah Vernon

4:19 PM

@Charlieface SQL Server

so, is 80 minutes wait for a shutdown a long time? No, because it still hasn't shut down. dba.stackexchange.com/questions/316949/…

Paul White

> If I request the SQL Server service to shutdown via SQL Server Configuration Manager, or via the Services control panel applet, or via net stop mssql$instance all logins are disabled except members of the sysadmin role...

That paragraph is confusingly worded

@HannahVernon If you issue a T-SQL SHUTDOWN (without NOWAIT) what happens?

Hannah Vernon

it shuts down immediately

I'll add that to the question

John K. N.

@PaulWhite They (person submitting a the pull request) want you (the owner of the repository) to consider pulling in their changes to your repository, thereby making your repository betterer.

Paul White

@HannahVernon Are local machine administrators members of sysadmin (or have CONTROL SERVER) permission?

Hannah Vernon

let me confirm, but yes I expect they are.

Paul White

4:27 PM

@JohnK.N. Yeah I get that but the action is taken from my perspective, not theirs

"push request" would make more sense

John K. N.

@SeanGallardy-MostlyRetired Sourcetree from Atlassian is a helpful GUI.

Erik Darling

wouldn't "change request" be best way to frame it

you're pushing, they're pulling

but it's all a change

Paul White

@ErikDarling Well yeah but now you're dreaming

John K. N.

@ErikDarling That might conflict with ITIL v3.0

Paul White

Like software would ever use plain English

@JohnK.N. WTF is ITIL v3.0

Sean Gallardy - Mostly Retired

4:29 PM

@JohnK.N. I was looking at merge vs rebase and came across their site - very useful, I'm just still such a noob I don't understand the downstream and ultimate ways it'll affect everything.

Erik Darling

@PaulWhite that explains why john is making shit up

John K. N.

Information Technology Infrastructure Library.

Sean Gallardy - Mostly Retired

@HannahVernon If you want to get a dump and share it with me? use procdump -ma

Paul White

@HannahVernon Check for permission errors in the Windows logs while you're at it

@JohnK.N. WTF is "Information Technology Infrastructure Library"

Sean Gallardy - Mostly Retired

@ErikDarling Like download and upload

John K. N.

4:30 PM

Developed by the British Govt. to bring some structure into IT

ITIL

The Information Technology Infrastructure Library (ITIL) is a set of detailed practices for IT activities such as IT service management (ITSM) and IT asset management (ITAM) that focus on aligning IT services with the needs of business.ITIL describes processes, procedures, tasks, and checklists which are neither organization-specific nor technology-specific, but can be applied by an organization toward strategy, delivering value, and maintaining a minimum level of competency. It allows the organization to establish a baseline from which it can plan, implement, and measure. It is used to ...

Sean Gallardy - Mostly Retired

and yet <made up statistic number here>% of IT is still wild wild west

Paul White

Is anyone here British

Hannah Vernon

@PaulWhite so my account, which I can use to login to the server is a member of the local Administrators group on the server machine. I am also a member of the sysadmin server role in the SQL Server in question. I can see in the SQL Server error log that a shutdown was initiated because the server writes that into the error log when I use service controller to shutdown the service.

Sean Gallardy - Mostly Retired

Aren't you?

Paul White

@SeanGallardy-MostlyRetired No

John K. N.

4:31 PM

@PaulWhite My passport says so

Sean Gallardy - Mostly Retired

I thought you were from England

Hannah Vernon

@PaulWhite I am technically

Paul White

@SeanGallardy-MostlyRetired I can kick you you know

Is anyone here currently in Britain
Pedants

Sean Gallardy - Mostly Retired

I was being serious, I thought you weren't originally from NZ. As per usual I was probably wrong.

John K. N.

wrong person

Hannah Vernon

4:34 PM

@PaulWhite I filtered the Windows Security Event Log by "Critical, Errors, Warnings" and nothing shows up. Filtering the Application log shows nothing related to SQL Server except the Report Server complaining that it can not log in to SQL Server. The System event log shows now errors or warnings around SQL Server either

John K. N.

@PaulWhite I wanted to go to show my respects. Then I saw the queues and chose a Negroni instead.

@HannahVernon did you stop and disable all the SQL * CEIP services?

Hannah Vernon

@JohnK.N. yes, the CEIP service is disabled.

which I realize is not technically a supported scenario

Paul White

@HannahVernon Anyway, how ever you have things configured, when you try to shut down from outside SQL Server, you don't get a permissions error.

Charlieface

@HannahVernon Starting to feel like a Stack Overflow comment thread this. What, pray tell, are the OS build numbers for the DC and the client machine? Are they a state secret?

Paul White

The BROKER_SHUTDOWN wait on spid 9 looks odd

Charlieface

4:41 PM

This is regarding the unrelated Kerberos issue, in case anyone is confused.

Hannah Vernon

@Charlieface lol sorry Charlie - I'm not a domain admin so I'm at a loss around the version details for the AD controllers.

@PaulWhite I wonder if DBCC INPUTBUFFER would show what that is running.

will take a look shortly

going for a smoke real quick lol

Paul White

@HannahVernon Do you see a log message saying the broker manager has shut down?

Erik Darling

just pull the plug

sheesh

in more important news, i have this extended event set up. it catches statements but not query plans. i'm getting salty.

or rather, it catches far fewer query plans than statements et al

Sean Gallardy - Mostly Retired

the offer to look at a dump stands

3

Paul White

@HannahVernon Also look in sys.dm_exec_requests. It's possible the server is trying to shut down cleanly, not waiting on anything, but actively doing stuff. IIRC not all broker stuff uses a session.

There are broker-specific DMVs as well e.g. sys.dm_broker_activated_tasks

But I feel sure you would have mentioned if this server were being used for something unusual like service broker

Hannah Vernon

4:51 PM

Hahahah

@SeanGallardy-MostlyRetired are you talking to me, or?

Paul White

22 mins ago, by Sean Gallardy - Mostly Retired

@HannahVernon If you want to get a dump and share it with me? use procdump -ma

Hannah Vernon

@PaulWhite somehow I missed that

Paul White

Probably your first day using chat

Hannah Vernon

probably

its definitely my first time using procdump lol

Paul White

Is it possible there are a large number of broker conversations active?

Hannah Vernon

4:59 PM

@PaulWhite very unlikely - Service Broker is not being used in this instance other than for the usual msdb stuff. I have another new VM that I'll install SQL Server into with nothing other than plain vanilla install and see if it happens there too.

Paul White

> the usual msdb stuff

Erik Darling

🙄

Hannah Vernon

ok perhaps I should clarify "usual" means i haven't added anything using SB.

Paul White

19 mins ago, by Paul White

@HannahVernon Do you see a log message saying the broker manager has shut down?

Is sys.transmission_queue empty?

Sean Gallardy - Mostly Retired

@HannahVernon Yep

Paul White

5:04 PM

Nuclear war can't come too soon

Erik Darling

with your luck you'd survive

Hannah Vernon

@PaulWhite "Service Broker manager has shut down" message was immediately recorded in the SQL Server Error Log following "SQL Server is terminating in response to a 'stop' request from Service Control Manager"

Sean Gallardy - Mostly Retired

Wouldn't all the climate people be happy? Wouldn't it bring about nuclear winter and thus not have temps raise by 0.1 degree in the next 100 years?

Paul White

@ErikDarling Aint that the truth

@HannahVernon Good so it's not SB then.

You should hire a consultant. I can recommend one if necessary.

Very reasonable rates, I'm told.

Erik Darling

not it 👉👃

Paul White

5:10 PM

Microsoft Support are also available

Hannah Vernon

interestingly enough if I start the instance without starting the Agent service, then it shuts down via net stop without issue.

I continue to investigate

Paul White

Sounds like an Agent problem

Sean Gallardy - Mostly Retired

@PaulWhite Good luck with that these days...

Hannah Vernon

@PaulWhite "yes, I've rebooted"

Paul White

@SeanGallardy-MostlyRetired What do you mean these days 😀

Sean Gallardy - Mostly Retired

5:12 PM

@PaulWhite You know what I mean ;)

Paul White

Yeah

Sean Gallardy - Mostly Retired

I'm actively helping the best people left in that org find new positions.

So far I've been able to place a bunch of people

literally hundreds of years' worth of experience are gone and not being replaced

Paul White

No problems with SQL Server Agent in The Cloud

@SeanGallardy-MostlyRetired Yeah well it couldn't happen to a better employer

Sean Gallardy - Mostly Retired

They have resorted to hiring people who failed all tech screens and previous tech interviews

Literally failing their way into a job

Hannah Vernon

> Dump 1 writing: Estimated dump file size is 14469 MB

should I disabled LPIM

Zikato

5:15 PM

Morning

Sean Gallardy - Mostly Retired

@PaulWhite I'm mixed on their employment. On one hand they pay horribly below every other tech place, they have slightly worse benefits, draconian policies, no hardware budget, and you're expected to leave to another company and come back to get any promotion. They hire the worst people to hit quotas who can't do their job functions.

2

Paul White

Zikato

I've missed out on a lot

Paul White

@Zikato Honestly, today is the day to skip the transcript if you can

Sean Gallardy - Mostly Retired

@HannahVernon Nah, you're good.

Hannah Vernon

5:16 PM

@SeanGallardy-MostlyRetired how do you want it, sir?

2

Sean Gallardy - Mostly Retired

On the other hand, there are some insanely and amazingly genius people that I get to work with and learn from, which doesn't make up for everything, but is something for my own personal growth.

Zikato

I agree, pull request isn't an intuitive name. GitLab is using a Merge request

Sean Gallardy - Mostly Retired

@HannahVernon Zipped and put it whereever, you can shoot me the link via email (if you still have my email)

Hannah Vernon

ok

Zikato

Also whitespace should be automatically resolved, IMO. I think it depends on your diff tool?

Paul White

5:19 PM

You're a diff tool

4

Hannah Vernon

flagged for mod

Sean Gallardy - Mostly Retired

popcorn.gif

Erik Darling

is there a mod named maude

Paul White

replaced by mauve because faster

Zikato

@ErikDarling about the filtered plan collection - how about USE QUERY_PLAN_PROFILE and query_plan_profile XE? Would be nice if that had the SSMS integration as well.

Erik Darling

5:28 PM

those are useless to me

Hannah Vernon

@SeanGallardy-MostlyRetired you've got mail

Sean Gallardy - Mostly Retired

@HannahVernon Cool, give me a minute to DL

@HannahVernon ok, done, you can delete

Hannah Vernon

cheers

Paul White

@ErikDarling You mean it doesn't collect plans for statements without plans, or something else?

Erik Darling

@PaulWhite just as an example, it'll log an update...from that runs for 1 second, but there's no query plan for it.

i'm guessing it's triggers or something that make the statement run for > 1 second but the plan for the statement is shorter or something

or something

Paul White

5:43 PM

I see

Erik Darling

do u.

Paul White

well I understand what you're saying at least

which is forward progress

Erik Darling

hm yeah

if i only collect plans and drop the duration down to 100ms, the updates show up

so something else is afoot

Hannah Vernon

ahem, when I said Service Broker wasn't running anything I may have lied. A little.

also, killing the SB session lets the server shut down just fine.

Paul White

OMG

Hannah Vernon

5:56 PM

I feel like perhaps I should just go away now.

Paul White

I'm going to murder everyone on general principle

Hannah Vernon

its pretty weird because the Service Broker is only used to dispatch DDL Event Notifications. I'd like to understand why that blocks shutdown.

and how to mitigate it

aside from nuking SB from orbit, I mean

Sean Gallardy - Mostly Retired

@HannahVernon Let me guess, [dbo].[capture_server_wide_ddl_events]

Hannah Vernon

also, this SB setup is used on other servers that don't exhibit the non-shutting down behavoir

@SeanGallardy-MostlyRetired lol you guessed it

Paul White

6

A: How to quickly shut down all elements of Microsoft SQL Server Service Broker?

The reason is slow is likely because those elements are locked with a SCH-S lock because they are in use, thus preventing your drop statements. The typical culprit is activated procedures that are running in the background. This can be quickly investigated by checking the Activity Monitor blockin...

Sean Gallardy - Mostly Retired

6:00 PM

Yeah I see it's still running and not shutting down

which is holding up the broker shutdown

Hannah Vernon

@PaulWhite reading that now

Sean Gallardy - Mostly Retired

00 ntdll!ZwSignalAndWaitForSingleObject
01 KERNELBASE!SignalObjectAndWait
02 sqldk!SystemThread::SignalAndWait
03 sqldk!SOS_Scheduler::SwitchToThreadWorker
04 sqldk!SOS_Scheduler::Switch
05 sqldk!SOS_Scheduler::SuspendNonPreemptive
06 sqldk!SOS_Scheduler::Suspend
07 sqldk!WaitableBase::Wait
08 sqllang!CServiceBrokerMgr::NotifyInstanceShutdown
09 sqlmin!DBMgr::ShutdownAll
0a sqlservr!SignalProc
0b sqlmin!StartUp::InitDBMS
0c sqldk!SOS_Task::Param::Execute
0d sqldk!SOS_Scheduler::RunTask
0e sqldk!SOS_Scheduler::ProcessTasks

Paul White

55 mins ago, by Hannah Vernon

@PaulWhite "Service Broker manager has shut down" message was immediately recorded in the SQL Server Error Log following "SQL Server is terminating in response to a 'stop' request from Service Control Manager"

Hannah Vernon

I know. And I wasn't wrong about that.

Erik Darling

this might be one of the hardest OPs chat has experienced

Paul White

6:03 PM

@ErikDarling It's a biggie for sure

Hannah Vernon

I chose to live in Infamy, ok

Paul White

I'm still not 100% clear on how SQL Server Agent factors into it, but luckily, I do not care

Hannah Vernon

"local factors" obviously

Paul White

@ErikDarling Have you consulted the extensive documentation

Honestly, extended events is just an absolute nightmare for strangeness

I'd be interested in the answer if you ever figure out what's happening there exactly

Erik Darling

well, there are two separate things at play

1. a stored procedure is logged at sp_statement_completed, where the entirety of the procedure takes > 1 second, but no individual statement in the procedure took > 1 second

2. a statement took > 1 second, but the plan duration is recorded at ~100 milliseconds. these were for updates/inserts to tables that have triggers that run and do a bunch of tiny things (and perhaps fire other triggers) that end up taking > 1 second.

Paul White

6:12 PM

nice blog

Erik Darling

but the common thread is a calling task with a duration of > 1 second, but no individual sub-task that takes > 200 milliseconds

so i'm real josh-darned annoyed

hi josh

Paul White

nice

computers are so bloody literal

Sean Gallardy - Mostly Retired

@HannahVernon Do something for me for a test, it seems as though there is no timeout value in your activated SP - can you add a timeout value and see if it shuts down?

Looks like that was session 42 at the time

Paul White

The answer to Life, The Universe, and Everything

Sean Gallardy - Mostly Retired

Hard to tell if that session started or last received data at this time: 2022-09-15 17:11:30.000

my math could also be off

@HannahVernon it's weird, it looks like it's waiting for receiving a result set. If this happens again, can you check select * from sys.dm_broker_queue_monitors to see the status of the queue? I wonder if it's sittining in the RECEIVES_OCCURRING state. If it is, yeah, add a timeout to the receive.

Hannah Vernon

6:28 PM

Will do

Josh Darnell

@ErikDarling -_-

Hannah Vernon

@SeanGallardy-MostlyRetired the state is DROPPED, with tasks_waiting = 1

@SeanGallardy-MostlyRetired adding , TIMEOUT = 1000 to the WAITFOR allows the server to shutdown.

Erik Darling

6:47 PM

@HannahVernon your question should be closed as too localized 🤗

Paul White

Tip of the iceberg

Unclear what you're smoking

5

Hannah Vernon

oh Paul, I'm sorry

Erik Darling

grovel! GROVEL!

Josh Darnell

@PaulWhite Underrated comment 😂

Zikato

The Heap™ - come for the help, stay for Paul's roasts.

9

Paul White

7:02 PM

😀

Hannah Vernon

I don't want to sound too defensive, but in the other instances where this Server Event Notification is set up, shutdown via the service manager is no problem at all.

Sean Gallardy - Mostly Retired

7:31 PM

@HannahVernon Yeah it's waiting for the task to cancel, but the task is sitting in a waiting state forever because there is no timeout in the receive, so it just waits forever. If it would get switched in from something then it'd see the task is aborted and close, but there isn't anything to switch it in, hence the timeout addition working.

Paul White

In fairness, I think it's unexpected that situation would prevent a clean shutdown progressing

Sean Gallardy - Mostly Retired

I concur, definitely seems like a b-word

Paul White

Opportunity for Product improvement

Sean Gallardy - Mostly Retired

OPportunity indeed.

Paul White

😀

Hannah Vernon

7:41 PM

@PaulWhite I certainly found it to be unexpected 😉

John K. N.

7:59 PM

See the transcript from The Heap from this point to there for an extended discussion and click here for the solution on how to allow Service Broker to shut down gracefully when tracking [dbo].[capture_server_wide_ddl_events] in a stored procedure. — John K. N. 30 secs ago

pun intended

Hannah Vernon

8:18 PM

aaaaand, it's gone.

Sean Gallardy - Mostly Retired

8:37 PM

I missed it but there was another chat made and I got it

I sent them an email

@DepthofField You should have an email

Paul White

8:57 PM

Wherever works for you. If you don't want it public, join the Heap chat and I can give you my email and you can send me a link. — Sean Gallardy - Mostly Retired 1 hour ago

Context for anyone wondering what that was all about

Sean Gallardy - Mostly Retired

I just like to cause problems, naturally.

Paul White

You just like to look at other people's dumps

Erik Darling

9:14 PM

i looked at a dump once. ended up in the fecal position.

Paul White

9:34 PM

Top men will get to the bottom of it eh

Who starred that