Due to a accidental config file check-in, a unit test wiped out our entire production database (did a drop, then re-created the tables). Trying to restore from backup revealed the recent backups were corrupted and the only valid backup we have is just over 1 year old.
What are my options for re...
I created a maintenance plan to back up two user databases and system databases over a month ago. I queried the msdb database to view the size of the backups using the following query.
SELECT
bs.database_name,
bs.type,
(bs.backup_size/1024)/1024 [BackupSize (MB)],
(bs.compressed_backup_size/...
So who's to blame, if the Pats were under-inflated but the refs approved them? How is this different from Rodgers trying to get away with over-inflated balls and the refs approving them?
Anyway, a source claims the balls were given to the refs slightly under-inflated (not two pounds!) and they expected them to top them off (which other sources claim is the ref's job), but they didn't, they just approved them. Lazy? Sting? Who knows.
@AaronBertrand I was thinking about the same - he is either a pressure magician or there were some dwarfs hiding there, ready with manometers and whatever
Anyway newer reports (e.g. on profootballtalk) say that only one single ball was found to be around 2psi short, the rest were all about 1psi short. More and more stuff coming out points to a combination of ref failure and weather than them intentionally cheating. I don't doubt that they try to skirt the low end of the psi range but I have very serious doubts that they intentionally made the balls that low.
So many people are still saying "well why didn't the Colts balls deflate too?" groan because Luck doesn't want his footballs at the low end, and they didn't play with fire - probably because they knew there was going to be an issue.
Can't wait to see what Goodell is going to do. If he's going to claim that the refs measured the balls in the proper range during pre-game inspection, he's going to have to prove it. Burden of proof is so far the opposite of current public opinion it's not even funny.
I find it both amusing and depressing how much thought, time, and energy has gone into dissecting this incident, which boils down to a couple of extra puffs of air in a ball used to play a game that grown men are paid millions of dollars to play and other grown people are paid millions of dollars to talk about. Meanwhile actual problems in the world go ignored.
...but seriously people are getting killed by police in Egypt for attending a memorial service, there is absolutely insane stuff happening with monetary policy in Europe right now, and all anyone in this country can talk about are how much air was in these damn footballs.
I'm trying to figure out how to filter out certain data types while leaving the record displayed if the value happens to be NULL.
This picture shows what I am hoping to accomplish. For Drive ID 652132, I would only like the record with EMET=3 filtering out the other equipment types/description...
I'll try to explain this as simple as possible, since it involves an ecommerce music site.
So I have an app website based on CakePHP which is an ecommerce site, we sell online Music.
The problem I'm having is that after someone buys an Album, the Genre rows of that album get Deleted, and then t...
What do you all like to use for database DDL change tracking? DDL trigger, audit, rely on the logs from a deployment system or source control tool, or something else?
Looked in the help section to figure out where to post a bug report, it's minor but still.
The candle icon for "Achievements" has a bug at least for my account where there is +12 rep for Stack Overflow and a +2 rep for DBA though no text to the right like all the other rows.
When I go to Stack ...
I have this procedure :
alter proc rel_manipula(@inicio datetime,@fim datetime)
as
select A.codigoins, b.nomeid ,sum(a.qtd), a.dtmov from com_InsusTrans A
inner join com_insus B on b.codigoins = a.codigoins
where (a.dtmov >= @inicio) and (a.dtmov <= @fim)
group by a.codigoins , b.nomeid, a.dtm...
@AaronBertrand I prefer the DDL trigger myself too. The problem I've run into is that DACPAC deployments will drop the trigger if it's not in the database project. :)
@JamesLupolt I like to use a system where the patches are actually prepared up front and tested in a smoke-test environment before they go anywhere near test or prod environments:
i. Prepare target schema ii. Prepare patch iii. Restore current version to smoke test environment iv. Apply patch v. Run a schema diff against the target schema.
Nothing goes into any test or prod environment without a patch script being prepared by someone.
AND (Substring(CONVERT(varchar,[FDFundingDate],112),1,4) = COALESCE (@Year,Substring(CONVERT(varchar,[FDFundingDate],112),1,4)))
AND (Substring(CONVERT(varchar,[FDFundingDate],112),6,2) = COALESCE (@Month,Substring(CONVERT(varchar,[FDFundingDate],112),6,2)))
AND (TitleUnderwriterGroup IN (SELECT VALUE FROM ServicingLibrary.dbo.StringParse(',',@TitleUnderwriterTitleUnderwriterGroup) WHERE VALUE = TitleUnderwriterGroup))
AND (DocumentType IN (SELECT VALUE FROM ServicingLibrary.dbo.StringParse(',',@DocumentDocument) WHERE VALUE = DocumentType))
what filters are used in the query in question though
month and year as strings?
if the query runs more than a couple of seconds you may benefit from an OPTION (RECOMPILE) on that query if it has a date filter that generates a crap plan
No clue man. I was just looking for new reports that are using this large table so that I can index it properly since it currently just has a PK that's on an auto incrementing integer.
I'm really annoyed by that damn convert why the fuck cares what the date is formatted like! Why not just compare the year and the month then god dammit?!
Also the testing data left here is nonsense....
@Month varchar(15) = 'WFDHPAOM',
^not a month on this planet.
Sorry for the rant there folks but this is some stupid code.
[dbo].[StringParse](@delimiter [nvarchar](1), @string [nvarchar](max)) RETURNS TABLE ( [value] [nvarchar](4000) NULL ) WITH EXECUTE AS CALLER AS EXTERNAL NAME [ServicingLibrary].[ServicingLibrary.Library.StringParser].[InitMethod]
I have this procedure :
alter proc rel_manipula(@inicio datetime,@fim datetime)
as
select A.codigoins, b.nomeid ,sum(a.qtd), a.dtmov from com_InsusTrans A
inner join com_insus B on b.codigoins = a.codigoins
where (a.dtmov >= @inicio) and (a.dtmov <= @fim)
group by a.codigoins , b.nomeid, a.dtm...
I refuse to. I get to say yay or nay on all of the ETL that comes through. If it's not up the snuff then I get to say fix it. I don't care about the deadline or any of that crap. If they want to go over my head fine but my sign off will not be on something that's crappy.
I searched the Help and the meta questions to see if this is something obvious. I've been reading several posts on the dba stackexchange and have so far come across 3 or 4 that have... Attachments? (See the attached "figure" in this post as an example.) And for each of them, it looks like the...
@Zane we are up against the end of our sprint and it's got to get to QA for validation. The code works but I know there are better, more performant (cc @AaronBertrand) ways to do it
People are really pushing it I ask a guy on a scale from 1 to 10 how bad he needs something he says it's not a 10 but it has to be done by noon tomorrow?!
I like hearing Belichick because he's so smug and doesn't bend for the media. Tells it like it is, doesn't answer questions he doesn't want to, and makes awesome and completely unexpected My Cousin Vinny references.
Most of the players are just towing the company line, yeah we're going to try hard, play our best, contain their best, yawn-fest. With a few exceptions, they're smart enough to not reveal anything interesting, because they know the other team is listening too.
I think the mayor made the right call. If he hadn't, and the storm was closer to predictions, and people got hurt or worse, those same fuckwits complaining that they had to be home by 11 would be calling for his head on a stake.
@ConcernedOfTunbridgeWells We have some variation of that for most apps at work. However, right now every app has a different deployment method, with some using change scripts, others using DACPACs, and others using various other tools. The apps using DACPAC deployments typically get the worst results and the weirdest unexpected changes compared to any other method.
@ConcernedOfTunbridgeWells The change script method has empirically been the most reliable. It's also the method that creates the most DBA monkey work (reviewing and testing change scripts written by 'agile' developers and arguing with them), at least until something goes wrong.
One way I do change control is to build test environments from the main data model. Then tests are run on that environment. Then you test that the change scripts actually change to that DB schema.
Let me rephrase that. You need a *competent* DevOps analyst
I think the current plan is to get a better audit or some DDL triggers in place shortly, so that we can prove what the deployments are actually doing on non-production environments before they hit prod.
Take your reference schema and the current version deployed. Run the patch script Do a comparison and check that the patch aligns the schema with the target.
@ConcernedOfTunbridgeWells I think the other condition to watch out for is when the new schema matches the target, but the target includes some changes that someone didn't realise were there.
For example, this weekend, someone accidentally deployed a broken version of a view as part of an unrelated deployment.
Then you test in the environment from the repository model and use the schema diff tool to check that the patch scripts align the current build with the repository.
<--Changing a 2 year old process to include $ signs in the numbers we give the the source system. Never mind the fact that 100% of our inventory is US Dollars.
@JamesLupolt The red gate tools are far superior to the Visual Studio ones. They're red gate's bread and butter, while Microsoft included rudimentary support to check a check box.
@AaronBertrand I couldn't fault them when I used them in the past. The change scripts they created worked too. I'll probably make another attempt at getting someone to pay for it this year. Perhaps the arguments you make in that blog post will help.
@JNK procedures, views, and functions - their definitions are so easy to pull. Changes to a table, on the other hand, are not very easy to generate from the metadata - those would seem to me to be more in PowerShell's wheelhouse.
@JNK sweet
iPhone revenue > (Microsoft revenue + Google revenue)
Think about that for a moment.
I remember when Bill Gates invested in Apple to prevent them from going bankrupt. Now Apple could buy Microsoft with cash
@JamesLupolt meaning ... you lose disk ... errorlog fills with errors, but AG doesnt automatically detect it and force a failover to its synchronous replica. similar to the situation described here. connmove.eu/en/…
just didnt know if anyone had any experience with it
yeah. we're doing an HA test with Pure and killed all the paths to one node, expecting it to detect catastrophic disk loss and force a failover. it didnt.
@JamesLupolt sure. we have AGs functioning just fine for the most part, but this is something that we just discovered. we have a 2+2 cluster (2 HA pairs, one of which is a DR pair) and we were hoping for automatic failover on disk loss
@JamesLupolt yep. single database in this AG
2 nodes in Denver, 2 nodes in NYC, and a fileshare witness in Dallas
What am I doing wrong? All the web says "to look in the Certificate provider to find certs with this dns name, use command Get-ChildItem -Path Cert:\LocalMachine\My -DnsName "Will we ever know?" and lo and behold, it ACTUALLY WORKS on most of my machines.
see: https://technet.microsoft.com/e...
@jcolebrand Nm, I'm encountering the same thing on a couple machines here, but none of my certs stored have a DNS listing. It could be that the parameter is only valid if one of your keys has a DNS name associated with it, but that's just a blind stab.
Due to a accidental config file check-in, a unit test wiped out our entire production database (did a drop, then re-created the tables). Trying to restore from backup revealed the recent backups were corrupted and the only valid backup we have is just over 1 year old.
What are my options for re...
I created a maintenance plan to back up two user databases and system databases over a month ago. I queried the msdb database to view the size of the backups using the following query.
SELECT
bs.database_name,
bs.type,
(bs.backup_size/1024)/1024 [BackupSize (MB)],
(bs.compressed_backup_size/...
I'm trying to figure out how to filter out certain data types while leaving the record displayed if the value happens to be NULL.
This picture shows what I am hoping to accomplish. For Drive ID 652132, I would only like the record with EMET=3 filtering out the other equipment types/description...
I'll try to explain this as simple as possible, since it involves an ecommerce music site.
So I have an app website based on CakePHP which is an ecommerce site, we sell online Music.
The problem I'm having is that after someone buys an Album, the Genre rows of that album get Deleted, and then t...
Looked in the help section to figure out where to post a bug report, it's minor but still.
The candle icon for "Achievements" has a bug at least for my account where there is +12 rep for Stack Overflow and a +2 rep for DBA though no text to the right like all the other rows.
When I go to Stack ...
I have this procedure :
alter proc rel_manipula(@inicio datetime,@fim datetime)
as
select A.codigoins, b.nomeid ,sum(a.qtd), a.dtmov from com_InsusTrans A
inner join com_insus B on b.codigoins = a.codigoins
where (a.dtmov >= @inicio) and (a.dtmov <= @fim)
group by a.codigoins , b.nomeid, a.dtm...
I searched the Help and the meta questions to see if this is something obvious. I've been reading several posts on the dba stackexchange and have so far come across 3 or 4 that have... Attachments? (See the attached "figure" in this post as an example.) And for each of them, it looks like the...
What am I doing wrong? All the web says "to look in the Certificate provider to find certs with this dns name, use command Get-ChildItem -Path Cert:\LocalMachine\My -DnsName "Will we ever know?" and lo and behold, it ACTUALLY WORKS on most of my machines.
see: https://technet.microsoft.com/e...
