« first day (4817 days earlier)      last day (346 days later) » 

00:43
"Moderate"
 
5 hours later…
06:12
A chairde - Morning all!
06:24
@HannahVernon No, not at all - that's what mods do! :-) You mentioned the word "gibberish" in your critique of my answer, which was in hindsight, harsh, esp. for relative newbies (OP 29 day member as of now, answer - New Contributor). I focused on that word but was blind to the bit at the top which you changed, which is fine by me - no probs with other bits either.
Too often I've seen words/phrases like "arrogant", "cliquish", "echo-chamber" & "inner-circle" when referring to (some of) the sites - I want this (these) site(s) to be welcoming to newbies (see here) . After all, nobody was born knowing SQL, well..., except perhaps for Erwin Brandstetter? I'll only use the words "gibberish" and "rubbish" to refer to comments here in future and not on the main site! :-)
06:39
Wordle 998 4/6*

⬛⬛🟨🟨⬛
🟨🟨⬛⬛⬛
⬛🟩🟩🟩🟩
🟩🟩🟩🟩🟩
@HannahVernon "as little as possible" is also good.
Had a good one at work today - postgres DB had "lots of shit" turning up in it.
mad panic, "have we been compromised" etc etc
war room meetings, people in other timezones dragged out of bed
much kerfuffle, then one of the sysads notices that the traffic is coming from an internal IP. ....
Guess what piece of corporate-helpware was the cause....
Fricking Nessus Scanner
that is the definition of Self-Pwnage.
(If it sounds fake its cos I'm glossing over a lot of detail that is irrelevant like an elephant)
 
4 hours later…
10:31
Wordle 998 6/6*

⬜⬜🟨⬜⬜
⬜🟨⬜⬜⬜
⬜⬜🟨🟩⬜
⬜🟩⬜🟩🟩
⬜🟩🟩🟩🟩
🟩🟩🟩🟩🟩
 
2 hours later…
12:40
@Criggie Oddly common: seangallardy.com/…
12:51
> ...maybe it’s time to rethink hitting your airplane with birds before it takes off to show how resilient it is to bird strikes. It’s stupid.
Maybe it is? Maybe it isn't?
At least consider throwing fewer birds less often
> First of all, do not use the join predicate to filter. Filtering is the RESTRICTION operation in the relational algebrae and is not a JOIN...
another sql professional chimes in
@Vérace awesome, thanks
@ErikDarling seems to think SQL is a relational algebra
13:02
Algebra gets around, lots of relations with her lately
> Your query properly writted
@SeanGallardy Why is the main (default) font of the text body so small?
From the same pro author
Did he Hit it first time?
@PaulWhite Properly irony'd.
13:06
I do believe he's French
@JohnK.N. 🤷‍♂️ I didn't write the CSS, but I can probably figure out how to make it larger.
@ErikDarling a real pro, of course (intonation going down towards the end)
@PaulWhite some think it's algebra, others think it's a paragraph
Yes, French
13:08
I was wondering what you were talking about, but now it makes complete sense
@SeanGallardy I found it uncomfortably small to read as well. Never used to be like that, I believe
Weird, ok let me check it out
Seens fine to me
You're reading on mobile tho
13:09
Yes I am
We're not
Maybe that's a clue for Sean
ugh
Chrome, non-mobile (aka desktop)
Microfiche
Other posts have larger paragraph font. It looks like the font size was overridden in that one post.
13:10
wow, yeah it's oddly small
7
let me go change it
that's what she said
good work everyone
@HannahVernon story of my life
That got starred so quickly
🤣🤣
Give the people what they want
By the way, on @Zikato's recommendation, I set up DBA Dash to monitor our SQL Server's at work. It has been pretty neat so far. So thanks, @Zikato!
Great work, @Zikato
I guess saying @Zikato recommended it might be overstating things.
Feb 20, 2023 at 18:59, by Zikato
I don't, but I've heard good things about https://dbadash.com/
@Zikato being @Zikato
Sounds like a recommendation from @Zikato to me
13:16
Also, I can't believe that was a year ago.
@JohnK.N. I changed it, would you hit f5 and see if it's better?
@JoshDarnell was it annoying to set up?
@SeanGallardy It's better for me, but I had to hard refresh to see it, even opening the link afresh
@PaulWhite Thank you :)
Don't thank me, thank @Zikato
13:18
I wish I knew a software company that made a product or could make a product that would help people write documentation and keep it updated.
@ErikDarling it was made by a wiseman and @Zikato recommended it, how bad can it be
TroubleShooting Guide
Ridiculous acronym
13:19
the query plan for the query the frenchman took offense at is quite funny
well something weird happened to chat
Fred le Frenchie
My chat is broken, I believe
Could be CoPilot error
Not a very good CoPilot
13:22
That's been happening quite frequently lately
I didn't see it this time around
Still don't see the missing messages
I see doubles, even tried a force refresh
Someone's using NOLOCK
Also tried deleting, but then it just acts like it doesn't exist
@PaulWhite We're safe here, we only use READUNCOMMITTED and never use NOLOCK
yes, i get page not found for the permalinks to both of my messages
@SeanGallardy uch etter
13:25
It's a mystery
well that's all amusing
I eagerly await Erik sending chat invoices
doubles all round
somewhere, an SO dev is wondering if anyone noticed what they just did
@ErikDarling No, it was fairly easy. There were some places where the setup guide could be improved, but nothing that would block being able to set it up.
Boeing should be doing more birds.
13:28
So should Sean
2
Basically set the destination server (where the collected data will live), and the source servers (where you want to collect data from), and you're golden.
I try to eat as many birds as possible
I've gotten through quite a few
@JoshDarnell is it full of memes?
@ErikDarling I'm not an expert, but I don't think so. At least, they aren't prominently displayed. You can get to the memes if you want them. I just dug around and found PLE in one of the tabs, but it's sort of listing all the performance counters.
13:36
Here's the "Performance Summary" tab for all the servers I'm currently monitoring.
200ms read latency isn't highlighted
All of that seems like useful info. Except maybe IOPs.
@SeanGallardy And 10.1 is 🤨 Yeah, that's weird.
@ErikDarling if you're interested in seeing anything specific, let me know!
@SeanGallardy Algebra - could I have her number? :-)
@Vérace I'd give it to you but she can be a bit irrational
Frédéric Brouard - his website is hilarious - "PostgreSQL ayatollahs"! I suppose that would make me a member of the Republican Guard at the very least...
13:51
@JoshDarnell is there anything for query performance? blocking? deadlocks?
@SeanGallardy If she "gets around", I don't care!
Hi all, if I have a question about SSMS connections to SQL servers being blocked by firewall/network settings, is that something I can ask about troubleshooting here on DBA.SE? Or should that be asked on Server Fault or somewhere else?
1
Q: Unable connect to SQL Server on SSMS and Got error 10054

DavinI have an issue with connecting to a server on SSMS SQL Server. I got this error 10054 when i tried to connect : SQL Server error 10054 =================================== Cannot connect to <IP>. =================================== A connection was successfully established with the server, but t...

I'd do a search first. Depending on your issue there might already be an answer.
If your situation is unique and contains enough information, then you can always ask on dba.stackexchange.com
@TylerH It's on topic, but you'd have to write the question so it is immediately answerable without 1373878378247 back-and-forth comments. Those don't work well as Q & A. As John said, in other words.
Thanks both--yeah, my situation is a little confusing. We have SQL servers configured that work fine (e.g. can be connected to) from Windows 10 machines. But new Windows 11 machines in our env are locked down and then opened up as needed via whitelisting policies (rather than the inverse for Win10 in our environment where we locked down specific things), and I can't connect to some SQL server environments from SSMS in Win11, with a connection error (MSSQL Server Error 5 - Access is denied).
I will definitely search the site before asking (not a stranger to stringent curation on SE ;-))--it's been frustrating as most threads on the issue cover scenarios where no machine can connect to the SQL Server due to it being new or incorrectly configured
14:05
Be sure to include the SSMS version and full error messages (including state number)
And anything pertinent about the SQL Servers, including any listener setup
@Vérace Well some of the moves in higher dimensions is quite amazing
Someone will suggest running SSMS as an administrator
Firewall Rules can be a thing in a corporate environment on corporate clients. (Advanced Firewall Settings)
Connection strings are another issue. (FQDN)
I would suggest, depending on the error, a packet capture from source and destination
Access denied is rather generic
14:12
I'll see you all later
I had that today too, with a 3 minute estimate
Don't connect to just YourServer, but instead use YourServer.CompanyDomain.CompanyTLD (e.g. server.company.com)
I dread updates now even more than before due to the chirpy Win 11 messages
Yep, all of these things I have tried. I am also thinking packet inspection or traffic inspection via Fiddler or something might be the next necessary step
7
Q: What are valid connection strings for SSMS login box?

John K. N.What are valid connection strings for SSMS login box? Every once in a while I will have to connect to an instance of SQL Server with some specific configuration setting or in a DMZ where no host names are available or I'll have to connect to the Dedicated Admin Connection (DAC) via SSMS or SQLCM...

shameless plug
14:17
:-)
@ErikDarling There is query performance stuff - an object execution page that lists procedures and functions and their execution characteristics (duration, cpu, frequency)
And a "slow queries" page that splits out queries by duration and CPU and lets you drill into their details.
UI design has really come a long way over the years eh
I'm surprised I don't see anything about memory grants in either of those pages.
There is something about blocking too, but not deadlocks as far as I can tell.
@PaulWhite GRIDS AS FAR AS THE EYE CAN SEE
Oh, memory grants are in the "Running Queries" page.
If you click on the rows here, it takes you to what is essentially the sp_WhoIsActive output for that time period.
Someone is not parameterizing their SQL.
Ooo a pie chart
When including the full connection error (including the state), should I include the full "program location" stack trace?
Don't want to throw in an extra 100 lines of code if it's just noise
14:30
No
If possible in text form, instead of images.
@PaulWhite Can't ALL be grids I guess.
Any stack charts in there?
There was one earlier for waits
Oooh, yesssss.
14:36
OMG, I must need more coffee. In double checking the named pipes setting on the server, I saw it was disabled; enabling it and restarting the SQL Server service lets me connect on Win11 now. Though I'm curious why I was able to connect to it for years on Win10 without issue. I could have sworn it was enabled yesterday when I checked...
rubber ducking ftw
@TylerH Just as a matter of interest, why not Wireshark?
indeed
@Vérace I'm not really a networking guy, so no reason. Fiddler is just the only thing I've used (and in an extremely limited capacity). I don't even know if Fiddler and Wireshark are complete overlaps in terms of functionality, to be honest, or if they do somewhat different things
Neither checks named pipes on the server, as far as I know
That's wild, and now I'm confused. Why would named pipes matter in that scenario? I know pipes support remote communication, but I thought TCP got used by default when connecting to a remote SQL Server.
14:47
Sounds like a question for Server Fault
@TylerH Me neither - when I ask a question about it, it's "monkey is told, monkey does" - betimes I have vague glimpses through the mist of all that, but they are far too fleeting! (for a deep understanding)...
@PaulWhite -_______________-
@TylerH I never use Named Pipes and on checking a recently installed SQL Server 2022 instance I found that the protocol is deactivated in the SQL Server Network Configuration settings for the Network Protocols.
@TylerH It could be that the SQL Server Management Studio is using a fallback protocol (Named Pipes) on the Win11 client, because it is unable to connect via the TCP protocol. (hunch)
This could turn out to be quite an interesting question.
15:17
> Generally, TCP/IP is preferred in a slow LAN, WAN, or dial-up network, whereas named pipes can be a better choice when network speed is not the issue, as it offers more functionality, ease of use, and configuration options.
You crafty little you, you.
Wordle 998 5/6

🟨⬜🟨⬜⬜
⬜🟨⬜⬜🟨
⬜🟩🟩🟩🟩
⬜🟩🟩🟩🟩
🟩🟩🟩🟩🟩
15:32
@JoshDarnell depends on client protocol configured order
I suppose it's a personal thing, but I've always found named pipes to be less trouble
Imagine Anonymous Pipes
Someone asked me yesterday if we had monitoring for our SQL Servers. I had to then tell them about how incredibly amazing and wonderful the SolarWinds monitoring that we use is. They laughed.
you can't really hear the sarcasm in my voice when I said that, but they did.
It's a funny name for a company
funny I was just thinking that too
15:43
A thing that could wipe out all the electronics on earth one day
Better than Uranus Winds I suppose
2
I wonder how many people actually use column sets
I've never seen it in the wild
xml addicts only
@JohnK.N. I thought it could be a TCP issue, but I added a firewall policy to allow all TCP/UDP traffic (first it was port 1433 and 1434 for TCC/UDP specifically, then all ports for the SSMS application). But... not being a networking/security guy, I don't know if that actually worked/was applied correctly
of course, when I test it in PowerShell with test-netconnection, the TCP connection attempt does fail, so maybe you're right
except it works on the Win10 machine, and the TCP connection attempt fails there, too... or, at least I thought it did
15:56
solarwinds123
ha ha ha
that's a blast from the past
no kidding
i was considering doing a run of solarwinds123 t-shirts and sticking a box of them by their booth at pass
i bet a bunch would get handed out before anyone caught on
you're such a jerk
LOL I would have loved to have seen that
15:57
marketing people being the ultimate muppets
or something like "solarwinds, easy as 123"
subtle
I don't get it
(the solarwinds123 joke)
Tell me it wasn't a password
@SeanGallardy It was
you don't remember their huge security breach from a few years ago
I remember there was one, but I didn't look into it
@PaulWhite ... I ugh... wow.
16:07
> Another interesting development in the search for how attackers could introduce the supply chain compromise was the discovery of a compromised SolarWinds password existing on a private Github repository from June 2018 to November 2019. An intern working for SolarWinds had set the password solarwinds123 on an account that was interestingly granted access to the company’s update server.
> It has not been made evident whether the breached password played a part in the supply-chain attack affecting some 18,000 businesses with the compromised version of the Orion security platform. However, it is difficult to dismiss the possibility of the compromised password playing into the supply chain attack at some point.
that intern is definitely an extern now
4
Ultimately, it doesn't even matter exactly how the breach happened
Everyone thinks it was the "solarwinds123" password
Except Sean, anyway
I don't even know what to say
Time for us to make a better monitoring system with a much stronger password, something such as P@ssw0rd1
password can only be 8 characters long
Take the 1 off then
16:12
password may not contain any special characters
password it is!
password may only contain the word password
VSecure could be another option
see above
wompwomp
16:17
password may only contain the word password
Except in France, where "lepassword" is acceptable
it is writted
16:46
it is wrotten
17:28
ok, folks, I need a list of awesome songs to code by. Any suggestions?
probably just a shitload of depeche mode
17:44
I feel like that is not a bad suggestion
might get old pretty quick though.
it hasn't gotten old in 40 years, why would that change now
i'm taking a stab at sampling perfmon counters, if anyone wants to run it and yell at me about what's wrong
18:12
@JohnK.N. Looks like your hunch is correct. This answer suggests it is specified to work that way: dba.stackexchange.com/questions/318855/…
Given that that question and this question both exist (with answers), and the nature of my problem was due to TCP port 1433 being blocked but enabling Named Pipes being a workaround, I'm not sure whether there's any value in posting mine--it could just get closed as a duplicate of those two together
18:38
@SeanGallardy - I don't mind asking this on main but perhaps you could confirm what I'm thinking? I had a colleague ask about suspending data between the primary of a Distributed AG, and the forwarder, for a period of a day or so. I think the only way to "suspend" a Distributed AG is to set the role for current primary node of the Distributed AG to "secondary", then when you want to resume data movement, you need to do a FORCE_FAILOVER_ALLOW_DATA_LOSS. Something like this:
    /* suspend data movement, typically in preparation for a failover */
    ALTER AVAILABILITY GROUP [SQLDAG]
    SET (ROLE = SECONDARY);


    /* "undo" the preparation for failover */
    ALTER AVAILABILITY GROUP [SQLDAG]
    FORCE_FAILOVER_ALLOW_DATA_LOSS;
I tested it, naturally, and it seems fine. I just want the tiniest confirmation that I'm not missing something obvious.
FORCE_FAILOVER_ALLOW_DATA_LOSS seems drastic, but I don't think it is since there is not really any data loss.
19:07
@HannahVernon I can't say I've ever had someone ask me about this or figure a need to do it
Why not go to the forwarder and suspend db movement on the databases?
I'm not sure if that'd even work
19:40
its a curious one for sure
I've never wanted to do it before either. My colleague that is doing it is trying to figure out a way to comply with stupidity from mgmt.
Imma test it over a day to see how badly it kills my Distributed AG
if only there were a way to sneaky apply differential backups to stupid ags to help catch them up
log shipping once again supreme
The role change is internal to that instance, it may affect repl (CDC included) and potentially things like log reuse
I've never really thought about it, trying to think what other items might interact with that specific check
I can tell you it 1000000% was not tested
isn't that just assumed
shots fired
20:23
Did someone mention me?
@SeanGallardy I can also tell you that it takes the primary AG databases into a state where they can't be read, so that's a bit of a deal breaker
its an effective way to disable access to your databases :-)
I mentioned @Zikato
ooops I did it again @Zikato
 
1 hour later…
21:56
@PaulWhite This was an interesting read, thanks!
@Zikato 🙋‍♂️
In case anyone wants to tell me why my thought process sucks: A question on queue tables and race conditions. :)
for the record, this is the error I got when attempting to pause the Distributed AG:
> The target database, 'XXX', is participating in an availability group and is currently not accessible for queries. Either data movement is suspended or the availability replica is not enabled for read access. To allow read-only access to this and other databases in the availability group, enable read access to one or more secondary availability replicas in the group. For more information, see the ALTER AVAILABILITY GROUP statement in SQL Server Books Online.
@J.D. you got lmgtfy'd by nbk
also, I smell a queue
@HannahVernon lol it doesn't even answer the question.
@HannahVernon Yup, and we all know what that means. :D
@J.D. naturally, it's a comment, lol
Yuppers lol
22:05
Was
@PaulWhite heh, thanks.
@HannahVernon Everything about AGs is designed to reduce availability
I guess the 5-second version of what I'm trying to ask, is if my adds and removes are atomic, and the data is self-contained in the instance of the process that's working on those rows, is it still possible to have race conditions with concurrency.
@PaulWhite so far so good then
> application process crashes in an unhandled way, leaving SerialIds stuck in the queue table
this seems wrong. Shouldn't the application just handle whatever is next in the queue table when it starts up?
I mean it is a queue after all
that way you could run multiple apps and it would be no problem. If one crashes, just start another and it does the work remaining in the queue
I always imagined you listened to Another One Bites The Dust by Queen on loop
As you destroy SQL Servers
22:14
hahahahahah
Hannah, Destroyer of SQL Servers
I like it
I haven't read the question, but being atomic has nothing to do with being serialisable.
Maybe I'll read it later
I don't think I have ever answered a question on mobile
I recently read something about using OUTPUT like that for a queue table. If I remember correctly, it can have concurrency problems, but I'm probably imagining it
presses the destroy button
They should probably disallow answers from mobile devices
answering on mobile is purely terrifying
Remus Rusanu used to have a long rant about queue tables and all the ways people tried to be clever with them
I think he used to recommend DELETE with an OUTPUT clause but I might be misremembering
22:24
Another good day in 2-core server land.
@JoshDarnell Numbers go up, that's good, right?
@PaulWhite perhaps this, which does indeed show that
hah
@Yano_of_Queenscastle I like it. That's good spin. "We're experiencing consistent growth."
Could you put that data in a grid instead?
Failing that, a pie chart?
22:29
Sad to say no
Best I can do is a line chart.
Something for a future enhancement, perhaps
I'll file an issue.
Apparently there is an existing one for the lack of deadlock details: github.com/trimble-oss/dba-dash/issues/67
 
1 hour later…
23:39
Deadlocks seem like monitoring tool v1 territory
Weird

« first day (4817 days earlier)      last day (346 days later) »