« first day (752 days earlier)      last day (4127 days later) » 
00:00 - 20:0020:00 - 00:00

12:10 AM
@Phil I saw the one about not knowing that was a thing. This can be much more valuable, if properly sculpted.
 
 
9 hours later…
9:36 AM
@JackDouglas why on earth have this one been migrated here?
0
Q: back up a postgres db and restore on another computer

ayounesiI've just purchased a new computer (mac OSX) and I want to continue developing using the same database I had on the old computer. I don't want it remotely done cause I dont want to keep the other computer on. I just want to copy the db and put it on this computer for development. I have a USB sti...

And a brave user here:
0
Q: Database design of Ledger Account and Balance sheet generation

Nitesh VermaI am building an application which has deposit and loan accounts management, for which i have to develop ledger accounts and then generate balance sheet. Being new to database design, i am having a hard time in designing the database model best suited. Is there any standard model for the databas...

 
@dezso that one looks ok to me?
@dezso what is the link in your comment supposed to point to? I can change it even if you can't now...
Yes, you can. First step is reading the documentation thoroughly. — dezso 4 mins ago
 
thanks, I have copy&paste problems from time to time
@JackDouglas I feel it very basic and a bit lacking
But the main point is why a question with an accepted gets migrated
@JackDouglas recreated the comment
 
9:57 AM
@dezso Because these SO guys sent it over: mu is too short, kmp, Gajotres, Jan Hančič, Aleksander Blomskøld
we can reject the migration by closing the question, but I don't think we need to. The 'too basic' thing is mainly for SQL, stuff we want to shove over to SO, not vanilla administration imo
I take your point about the accepted answer
 
Morning
 
...ning ...ning ...ning ...ning
 
@Phil this is a bit sick - but how did you find this only 7 hours after uploading (or whatever you call that with GitHub)
 
@dezso My RSS feeds are vast and varied
 
@Phil so you get a notification about every new project?
 
@dezso No, it was linked from one of the geeky sites I read
 
@Phil ic
 
10:57 AM
 
 
1 hour later…
12:01 PM
@mat Heh, was just writing up a DETERMINISTIC thing for that Q
 
 
2 hours later…
1:35 PM
1
Q: MySQL Database Connection from PHP

John SecurayI have a rather simple issue, I'm trying to setup a website. I have created a mysql server for it, however when I go to the site I get this error. Warning: mysql_connect() [function.mysql-connect]: Can't connect to local MySQL server through socket '/tmp/mysqld.sock' (2) in /homepages/38/d4...

VtC off topic...i don't think the user is coming over anyway
 
2:20 PM
@DTest also done
 
3:09 PM
0
Q: Is that a good idea to use one database for 50.000+ shops?

faridmovsumovI know Shopify use only one database for all shops. But how can they handle their database with such big data? Is that a good idea to use one database for 50.000+ shops?

 
3:31 PM
@Phil he should probably use something web scale like us :(
 
@Phil let's add that one must be careful regarding data ownership
This does not provide an answer to the question. To critique or request clarification from an author, leave a comment below their post - you can always comment on your own posts, and once you have sufficient reputation you will be able to comment on any post. — dezso 35 mins ago
How does this work? The answer showed up in the first posts and low quality posts queues. In some order I DVd, flagged as not an answer and recommended deletion (IIRC w/ no comment)
And an RTFM alert:
0
Q: Using a Select in a Where clause for another Select

Stephane RollandI have made a draft remote application on the top of libpq for PostrgreSQL. It behaves well, but I have profiled the general functionning of the application. For each final business result that I produce, it happens that I call something like 40 select clause (over tcpip). I have reminiscences f...

 
Did SO/SE just die for a shot while?
 
@dezso It should probably be moved to SO as too basic
 
@Phil it did for me
 
3:42 PM
@swasheck oooo dream gig there.
 
but just an instant
 
@bluefeet VtCd as offtopic
 
Right now I'm pursuing 3m pretty heavily.
 
@Zane probably have to move to C - harlot
 
@swasheck They need to start writing Oracle tools & hire me ;P
 
3:44 PM
@Phil indeed.
 
@Phil I imagine the heapers swarming SQLSentry
 
@dezso next they'll monitor postgresql
 
@swasheck I'm unavailable for a while, sorry :)
 
4:14 PM
centOS is a huge pain in the ass
 
@Zane in what way?
 
@Zane you should try gentoo. much easier
 
hah
 
In every way lol. I'm used to having to do this much dicking around to change my domain name
 
4:27 PM
@Zane i'm an arch user, myself
 
I'll have to build a home box.
 
he has moved on to excel problems
0
Q: Macro Error in Line of COde

AndrewI have a macro that adds zeroes to every number in column a until there are 7 total digits in the number, it has worked fine up until today and I am getting an error in the line For1= i=1 To endrow = -1 and I cannot figure out what it means. THe code is Sub AddZeroes() 'Declarations Dim i As I...

 
@bluefeet no, i think that Andrew has macro-problems
 
I think I need to get a lappy from work I can install Cent on so I can get used to it.
 
Can't just build a VM for that? Use VMWare/Vbox?
 
4:40 PM
@MikeFal stop making sense
 
@MikeFal on my home machine I think not.
 
Why not @Zane? I've got a personal laptop that I run a couple VMs on for experimenting with SQL and Server 2012
 
@AaronBertrand "... 50,000 wingnuts?" for some reason this caused me to grin.
 
Problem is 2 fold. Problem numero uno is my laptop is garbage. Problem numero dos is if the company wants me to spend the few off hours I get learning things so we can switch from MSSQL to hadoop they are going to provide the hardware for me to do it.
 
probably dumb question: is there a popular linux desktop? Last time I ran a gui on a linux server it was between gnome and kde....
 
4:46 PM
@Zane Fair enough. If it's something the company wants you to do, they should provide you the resources.
 
and this was a decade ago or so :)
 
I'm guessing they're going forward with the whole Cassandra thing?
 
Oh yeah.
Actually they want to do PIG
 
@Zane so now it's hadoop? you can backend hadoop to sql server ;)
 
@DTest I couldn't tell you what the shell is, but I've used the Ubuntu distro and the desktop that is included is pretty easy to get around.
 
4:48 PM
@MikeFal Unity
@DTest what distro are you running/would you like to run?
 
@swasheck @MikeFal quick google says unity after 11.04, gnome before that
 
@DTest indeed.
 
@swasheck Am not sure yet...all I was told is I would be getting a linux laptop :)
 
@DTest depends on what you're going to do with it? if they're buying then i guess ask for RedHat + support :)
 
Yeah I'm about 75% sure they're going to do RHEL
which is fine with me, since I've used it...can 't say the same for ubuntu
 
4:51 PM
@DTest i'm confused, though. they're sending you a "linux laptop" ... why don't they sent it preconfigured?
 
@swasheck they very well might....but they might not preconfigure it the way I want it :)
 
@swasheck lol
 
not that I really know what I want, in regards to the gui ... so I guess it really doesn't matter
 
@DTest you're one of them thar mac folks, right?
 
yup!
 
4:53 PM
@DTest are you a fan of OSX because of the pretty, shiny interface? or because it's just *nix-y enough to get by?
 
Some things about the interface are neat (like the spaces discussion yesterday), but mainly I like a) "*nix-y" enough to get by and b) integration with gadgets (itunes + iphoto primarily)
 
@DTest mm. well then you won't like my rec. because i'm a big fan of minimalism so i split time between Awesome and OpenBox
 
@swasheck never heard of either...I'll check them out
 
It's also (a) a much less common target vector for sh*tware of all kinds (b) I find the browsers on Mac OS much more stable in general (and the first thing I do on every Windows VM I build is download Chrome) and (c) I
I'm even more bitter against Windows now because of what they did to Windows 8 to make it completely unusable. I laughed constantly during the lunchtime presentation we had on it yesterday - just about everything was "oh yeah, you have to learn this new keyboard shortcut to do that now"
 
@DTest OpenBox gained a lot of traction with the whole OLPC thing
 
5:00 PM
@AaronBertrand I wonder how true a) will be in the next decade for apple..I always thought as they gained market share, it would become a problem.
 
Mac OS is not identical to what it was 10 years ago, but even these innovators know better than to do what Windows 8 did.
 
@AaronBertrand the only reason i'd go to 8 would be for the native hyper-v (which i think i've said before). @DTest was asking for a Linux WM/UI and I was just trying to get a feel for what he liked. Many of the other features that he noted are independent from the WM itself.
 
@DTest they've gained plenty of market share and the situation hasn't really changed. And you'd think with the number of Android weenies who hate Apple fanboys there would be plenty of stuff out there...
 
I wanted a surface pro until I attempted use of windows 8
 
@swasheck yeah, I know. I use it as a guest only, so irrelevant there, and if I needed a Hyper-V host, I would probably run server.
 
5:02 PM
@AaronBertrand tried that too ... 2012
 
How many days until OSX is sandboxed inside iOS as "classic mode"?
 
@swasheck right, but you could have a Hyper-V host with Server 2008 R2 (which is whatI'm saying I would do)
But I've found VMWare generally (and perpetually) a little bit ahead of Hyper-V anyway.
 
@AaronBertrand yeah - i was just reading about the gains that Hyper-V made in 2012 and it was intriguing
@AaronBertrand apparently Hyper-V is better at managing IO than VMWare (but nothing else)
 
@swasheck would any of them really help a workstation running a Hyper-V host in Win8?
 
@swasheck openbox might work...This is generally what my work environment looks like anyway: openbox.org/wiki/Image:Openbox_dust_carbon.jpg
probably blocked for you though :D
 
5:05 PM
@swasheck I mean, really, are you using your client (not server) operating system for Hyper-V guests that are running things that are dependent on I/O performance? I think the last problem there is your choice of virtualization technologies.
Besides, I don't know how much that I/O stuff is going to matter once we're all on SSDs everywhere. I/O is quickly moving from biggest bottleneck to nearly irrelevant.
 
dba.stackexchange.com/questions/33541/… @AaronBertrand Great answer however with no source control at all OP is still probably not having a good time.
 
@Zane Everyone keeps giving me crap for shooting down source control, and I said multiple times on that page that it isn't a substitute for source control. It's a meantime-patch. Should I make that more clear in my answer so people stop telling me that they guy should be using source control? I never said he shouldn't be.
@Zane Even with a proper version control process in place, the DDL trigger will still call out events that happen when people go outside the process (which is sometimes a necessity).
 
@AaronBertrand I'm I'm not giving you crap for it.
 
@Zane several people have (including comments that have been deleted)
 
@AaronBertrand I said great answer. I'm saying at the shop where source control is not an option the OP is proabably having a bad time.
 
5:09 PM
@Zane it still sounds like you're telling me that the guy needs source control. I know this already.
@Zane probably, but you can't fix that with code.
 
Oooh, the snow's starting again. Apocalypse incoming!
 
@Phil So you should go buy bread now huh?
 
@jcolebrand I live above the village shop :P
 
I know. I'm just suggesting that his situation is a bummer. That must suck.
 
@Phil that's convenient
 
5:11 PM
@Zane yep. With a DDL trigger though, he can now show (whoever makes decisions about process) what chaos it is.
 
ha!
 
Truth. I have a similar problem in the DEV and QA environments here.
 
Full of cowboy coders?
 
The VP of development made a generic Sys Admin role and gave it to all developers.
So if someone screws something up and I go to investigate int just tells me the change was made blahblahAdmin
 
he likes to yell
i WANT THE ABOVE ANS — Ashish Ginotra 20 mins ago
 
5:14 PM
@AaronBertrand I just like to push my boundaries. In the end performance isn't all that big of a deal, other than to satisfy my own curiosity. It's also why I'm back to Win7 and VMWare Workstation
 
After next release I'm deleting that user weather the like it or not. If they want access they have to come through me. Muahaha
 
@DTest yeah. so basically OB is quite hands-on.
@DTest you have to pick a "dock" and i'd recommend feh as a wallpaper manager. Also, even if you're getting RHEL then I'd still recommend checking out the Arch Wiki as it has some really good documentation surrounding the underlying applications
 
Making a constructive observation about another persons answer is an easy way to get downvoted.
 
pfft, ssh is all you need to admin a Unix box. (With a Mac as a client, of course)
 
Grrr....what part of "Hold on, I'm validating things" doesn't make sense? I feel like I'm driving a car with 5 year olds. "Is the server ready yet? Is the server ready yet?"
 
5:22 PM
5864 µs = 5.864 ms right?
 
@Phil indeed
 
The average human eye blink takes 350 000 microseconds.
 
Bethany Marzewski on January 25, 2013

In December, we launched our 3rd annual Stack Overflow Annual User Survey to learn more about our site demographics and user trends throughout 2012. Compared to last year, we received an even larger sample size this year with almost 10,000 respondents!

Here are a few larger trends we’ve observed over the past three years:

You like us…you really like us!

Since 2009, site traffic to Stack Overflow has grown by a whopping 261.7%! As if this weren’t enough, we’re also now the 86th largest global site, according to Alexa. Our crazy goal of breaking into the top 50 is looking less crazy! …

SQL -> #2
 
I'm not entirely sure the backstory behind this: sharpsuits.net but ... I like it? (re: Not sure what the backstory is: Ireland's creative community got together to release a lot of pent up anger and sadness through the medium of the A3 poster, all in aid of Temple Street Children's Hospital.)
 
@MarkStorey-Smith :-)
@Zane where did you get down-voted? I saw the down-vote on your detailed answer involving SSIS and all kinds of screen shots, but that happened two weeks ago (not that I understand that down-vote, though).
 
5:39 PM
Must have gone away.
 
@AaronBertrand do you have any insights on this question?
2
Q: Warning in query plan "Cardinality Estimate"

Mikael Erikssoncreate table T(ID int identity primary key) insert into T default values insert into T default values go select cast(ID as varchar(10)) as ID from T where ID = 1 The query above has a warning in the query plan. <Warnings> <PlanAffectingConvert ConvertIssue="Cardinality Estimate" E...

 
It was on that clear execution plan answer. I got bumped to 0 but I'm back at 1
 
find it interesting
 
+ 1. Now let's start reading the answer :-). — Marian 2 hours ago
that's some street-cred
 
@Lamak I don't, I know the warning is new in SQL Server 2012, but I don't know whether it just does that for all casts/expressions (or all casts/expressions against columns involved in seeks/filters/sorts etc).
@Lamak I can do some digging.
 
5:44 PM
@swasheck if it's that long, it must be good!
 
Wow you can have 15,000 partitions good lord.
 
1
Q: How could i make a server side application which stores data in client side database

user2011819I am building an asp.net website which needs a database, but the requirements are such that i shall not store the data on the server. So i need help on how to store the data of a server side asp.net website on a client side database file. Means a web app interacting with a client side database fi...

 
@AaronBertrand I saw the closed connect Item that Mikael posted, and I find it weird that the cast could change the Cardinality Estimate
 
@Zane the default max is 1,000. They added support for 15,000, but you have to turn it on manually.
 
People that use things like hourly partitions need shooting
 
5:46 PM
@Zane they did this for SAP IIRC. They wanted to be able to partition by day, but the 1,000 limitation cut them off < 3 years.
 
@AaronBertrand And two upvotes already... I think the phrase rhymes with clucking bell.
 
I want to ask, "why is that the requirement?"
 
We partition by companyID in sets of 10
 
@Zane For archival or performance reasons? Sounds odd
Most partitioning is pointless because people don't understand what it's for and how it won't help if the app is indexed correctly.
 
Performance.
 
5:49 PM
I've mainly only ever seen it used for data archival/rotation purposes
 
@Zane how so? Is CompanyID the leading key in the clustered index? If so I don't think your partitions help anything.
 
Indeed it is. Everything in the datawarehouse is clustered first by company and then store.
After that it uses the natural key items of the table.
TransactionID, ItemID things of that nature.
 
Ok, I'm failing to understand how a partition that contains only companyID 1-10 helps a query for companyID 5
 
But after much testing with and without partitioning we found report query performance in an apples to apples scenario improved when we had them chopped into smaller partitions.
 
Maybe it helps distribute write performance but really that will only help if you have partitioning across filegroups and different I/O subsystems.
@Zane what kind of "report query"? For CompanyID 5, for companyID 1-10, some other subset, or all companies?
 
5:55 PM
a report query for company 5
 
A seek or range scan for a particular companyID should not get any faster because it's in its own partition. Now, if you had bad queries that had full table scans...
 
Nowhere in datawarehouse are we doing full table scans.
Is my understanding of partitioning wrong. Isn't it essentially the equivalent of having a database with only those 10 companies
 
@Zane then I still don't understand how partitioning has helped your report queries. <shrug>
@Zane yes I suppose, but if the companyID is the leading key in the clustered index, then SQL Server already has the ability to eliminate all of the pages in the clustered index that don't contain CompanyID = 5 (which is all of the other partitions PLUS 1-4 and 6-9 in the current partition).
Or 6-10 rather
 
6:18 PM
@AaronBertrand Question has (9 or 10) NOLOCK hints. First 3 answers, too :)
1
Q: Converting IN and UNION to something more efficient ( Probably EXISTS)

QuantbuffSELECT * FROM sec WHERE sec.cu IN (SELECT s.cu FROM sec s WITH (nolock) WHERE EXISTS (SELECT * FROM hact h WITH ( nolock), JOIN dport p WITH ( nolock) ON h.ic = p.ic ...

 
I saw that (and +1'd your comment). Wanted to suggest RCSI but they probably won't get it, so I moved along.
At least guffa didn't persist the nolock
 
If you were needing to use several no lock's why not just use read uncommitted?
 
Our database is very transaction heavy so we do not want our select query's to slow down our internal application. We dont mind dirty reads but we want faster application and no deadlocks — Quantbuff 15 mins ago
"We don't mind dirty reads" just always makes me flinch. I'm not an expert and I've not been around for that long, but I've really not come across a situation where the risks associated dirty reads are acceptable in production.
 
@AaronBertrand Yes and his suggestion looks much cleaner.
@swasheck I'm planning a game site at the moment. Not financial, just fun. I would not think of using dirty reads.
 
I use an RDBMS that doesn't do dirty reads, so any mention of it makes me shudder ;)
 
6:27 PM
@Aaron: The question made me wonder if SQL-Server always optimizes IN (SELECT ... UNION SELECT ....) the same as IN (SELECT ... UNION ALL SELECT ....)
 
@ypercube what kind of game?
 
Word games (Hangman and similar)
 
@ypercube nice. is there an educational aspect to it, or is it just for fun?
 
@ypercube I think it would have to be different - from the outside looking in, the UNION eliminates duplicates and therefore must introduce a sort. Though, I'll admit, I haven't tested this, there could be optimizations there that I don't know about. I wouldn't use IN in the first place...
 
@AaronBertrand someone should write a blog post about IN
 
6:31 PM
@swasheck I have one in the queue as a follow up to my LEFT JOIN / NOT EXISTS post. But it needs a lot of work and I suspect you'll see 5-10 other posts before I finish that one.
 
@swasheck as long as EXISTS and NOT EXISTS gets a mention too
 
@swasheck The final goal is to have an educational aspect, yes.
(through the fun).
 
@Phil what is this EXISTS thing of which you speak??? i like scans. they give me time to go get coffee. tablescans are like the DBA's version of "compiling"
 
6:34 PM
@swasheck there is this one by quassnoi
 
@swasheck Like the MySQL question someone mentioned yesterday? With the join of 60 derived tables?
 
@ypercube didnt see that one
 
I've said it before and I'll say it again... use of IN or NOT IN (SELECT x...) makes my skin crawl.
 
> However, NOT IN is different, since negation of NULL is NULL as well.
@Lamak interesting
 
@swasheck I like this explanation better: Since NULL is unknown, it is unknown whether NULL is equal - or NOT equal - to any other value, including NULL.
2
 
6:42 PM
@AaronBertrand I went cross eyed for a moment there
 
@AaronBertrand right. that's the "why" behind what quassnoi said, correct?
 
@swasheck right.
 
@AaronBertrand whew.
 
@swasheck but it's not as logically stated, IMHO. I don't think it gets the point across to the person coming across their first "why does NOT IN return the whole table???" experience.
 
@AaronBertrand i can see that.
 
6:49 PM
0
Q: Dates, times and SQL jobs

TomThis is probably a silly question but if I want a job to run Sunday at midnight would I choose Sunday 12:00AM or Monday 12:00AM?

</TROLL>
 
1
Q: mysql big query

Jovo KrnetaI have a complex app that solves problems of multilateral compensation , so i have up to 60 queries run - depending of the level of how deep the queries go for the level of 5 the query is like this : SELECT k.idpred1, k1.idpred1 AS k1idpred1, k2.idpred1 AS k2idpred1, k3.i...

 
Though he probably doesn't see why the question is hilarious in itself
 
@Phil what should i flag that thing as? irreparably stupid? my first-grader knows this
@ypercube that's sweet.
 
@Phil But what SE site should one use to ask questions about the mechanics of the modern day timepiece?
 
@MarkStorey-Smith VtC ... migrate to metaphysics.se
 
6:53 PM
Struggling to resist making my comment to the question just ever so slightly sarcastic
 
@MarkStorey-Smith flagged as too ignorant for this site
 
@MarkStorey-Smith Flagged as too complicated for a toddler
 
@Phil (barely)
 
Can I drop a link to my question in chat to get it some more attention? :)
 
7:00 PM
sure
 
2
Q: How can I fix "Cannot perform a differential backup ... a current database backup doesn't exist"?

RachelWe recently switched to the FULL recovery model, with full backups being done every weekend, and differentials being taken every day. The problem is, the differential backups don't always seem to work. The error message recorded in the job history is Executed as user: username. ... 9.00.3042...

 
@Rachel i was just getting ready to do that for you
 
Posted 2 hours ago and only 15 views >.<
 
@rachel Still having that issue?
 
@MikeFal related issue
 
7:01 PM
I'll look at it shortly, apparently I missed a rollback script request at my real job.
 
@MikeFal Yeah, I've accepted the fact that I probably shouldn't be modifying the system tables to specify the backup flag, so my 2nd question is to just ask how to fix the problem
@swasheck Thanks :)
 
as an aside, is there a way to dump the scripts for the maintenance plans created by maintenance plan wizard?
 
@Phil consider your hand slapped for leaving a flag to edit out part of an answer. The rest of the flag was valid.
 
@swasheck I'm not sure "scripts" are available, they're SSIS objects.
 
@MikeFal O_O
 
7:03 PM
@jcolebrand e-mail address removal?
 
aye
 
:P
 
anyone can edit
 
@swasheck I could get you the approximate T-SQL that gets ran from the Maintenance plans
 
You could also slap a profiler trace on it
 
7:05 PM
Oh, yeah. Obviously wasn't thinking :P
Must've been watching kitten videos on youtube at the time
 
@MikeFal Yeah I could do that too for the diff backup. I don't want to start a full backup during production hours though
 
@Phil meh, youtube, bleh
 
@ypercube totally random question: when you see words in your native language, do you "parse" them? for example, do you see genitive objects or just a word that you implicitly understand in such a way?
 
@Rachel Oh, sorry, the trace was more for @swasheck.
 
@swasheck Not sure I understand what you mean.
 
7:07 PM
For your issue, I don't think a trace is going to tell you much.
 
@jcolebrand Been installing SQL Server 2012 and an Oracle 11.2 RAC VM this afternoon, so my brain has melted with boredom
 
@MikeFal @Rachel i've just never used them so i didnt know what happens
 
@swasheck Gremlins eat your soul, that's what happens.
 
@ypercube nevermind. i thought i had a legitimate question but in thinking about it a bit more it was almost as ignorant as not being able to understand which day midnight is in
@MikeFal done.
 
1
A: Can I export a maintenance plan without using Integration Services?

billinkcMaintenance plans are stored in msdb.dbo.sysssispackages like any other SSIS packages that are stored to SQL Server. I have a handy post on SSIS Package Extract from MSDB that should cure what ails you.

 
7:09 PM
@rachel Not sure what specifically cases that issue, but have you considered upgrading to 2005 SP4?
 
@billinkc you said bacon
 
@swasheck Beggin' strips but DBAs will never know the difference
 
@billinkc The hell we won't
 
@MikeFal Its not an option for us at this current time, although it is being discussed for some point in the future
 
7:17 PM
@Rachel I understand the hurdles of applying SPs. :) I'm doing that very action tonight, in fact. I'm going to post an answer to the effect of that (with some other thoughts), but we can chat it out here if that works better.
 
@MikeFal Failover clustering has been a lifesaver
2
A: "Cannot perform a differential backup for database "myDb", because a current database backup does not exist."

HolocrypticAre you running VSS backups on the server (NTbackup, BE, etc.) for stuff other than the SQL data? According to this, the SQL differential backups can be invalidated if a backup software uses VSS for snapshots. As far as I know, LS does not use VSS just like Nick said. According to this thread ...

 
@swasheck Amen to that. I'm rolling out 2008R2 SP2 tonight and I will only have about an hour or actual app downtime as I roll across my 6-node cluster
The downtime comes from my app support guys preferring to shutdown apps gracefully before a failover, then restart.
 
@Rachel you should see if the SAN admin is running SAN snapshots. Those usually use VSS of some sort. They need to exclude your SQL Server files if they are.
 
(Because we have some terrible apps where the devs didn't properly code for loss of database connection underneath)
 
@MikeFal natch
@MikeFal one hour total, right?
 
7:21 PM
@swasheck Correct
 
@MikeFal how long do your failovers take?
 
We have 10 instances that support ~20 applications.
 
(per fail)
 
Most of them take ~2 minutes for everything to move over.
I have one instance that takes 10, simply because it's the instance that hosts 1100 databases.
 
@MikeFal sweet. tense 10 minutes?
 
7:22 PM
Can I just reiterate how wank the Windows 8/Windows Server 2012 GUI is?
 
@Phil Different. Takes getting used to, but it's not that bad.
@rachel Also, have you ever tested one of the successful fulls that your system tables aren't recognizing?
 
@MikeFal good point
i just got a recruiter email from a Unix admin post in Detroilet. Anybody interested?
 
@swasheck In a word, 'No'. In two words, 'Hell no'
 
@cfradenburg Yes, I'm getting a bunch of extra records other than the full backups, although there doesn't seem to be a very consistent schedule to them and the BackupPath is a GUID. In addition, the BackupSize is extremely small. I checked with our network admin and he said he thinks its LiveVault's volume shadow copy, and he'll disable the job for a few days to see if that makes a difference. — Rachel 3 mins ago
 
@swasheck Thanks, I've gotten a few comments on my question about that and lead me to asking our network admin about that. Apparently there is a 3rd party backup software that uses VSS, so I think that might be the culprit
 
7:27 PM
@Rachel he needs to exclude anything SQL Server-related.
 
@swasheck Oh, and no, not a tense 10 minutes. Just a long 10 minutes. Seriously, whoever designed this app to be client sharded should be abused with a frozen trout.
 
@Rachel aw come on ... he linked to the same question i did :)
he linked to the q - i linked to the best answer :)
 
@rachel As I'm reading the comments, the VSS things make sense. I retract my suggestion that SP 4 would fix that (but you should probably apply for other reasons. :D )
 
@MikeFal my 1min failovers are tense because i have manf. managers breathing down my neck
 
@swasheck We love you, man. You're a unique, perfect snowflake.
 
7:30 PM
@MikeFal just like everyone else
and i'd rather be a star
 
@swasheck lol I completely missed your comment, sorry
 
@Rachel no worries. i was just kidding.
 
@swasheck Man, when management does that, I tell them to go away and get the devs to clean their mess up. I have no tolerance for bullshit.
 
@swasheck There, comment updated accordingly :)
 
Unfortunately, it sounds like their clean up will be to refactor the database backend to be consolidated......on Oracle.
Oh goody, another person who doesn't understand recovery models.
0
Q: Transaction log for sql server is too big

gavinI have a database that has a 33 gb transaction log. It is caused by the fact that the database is in full recovery mode has been used for over 5 years prior to the first backup. I want to truncate and shrink the log file with the following command Backup log dbname with truncate_only go DBCC s...

 
7:32 PM
@MikeFal well those folks are about 3 miles away and they've caught on to my "buffer" time. I say that there's normally going to be about a 5 minute outage ... they try at 2 minutes and if it's still not up they start calling.
@MikeFal or the deprecation of Backup log dbname with truncate_only
@MikeFal debating whether or not to give him the gun with which he may shoot himself in the foot
 
@swasheck I will politely correct him.
 
dump tran with no_log Ah, the Sybase days
 
Plus, I'm shooting for 1k DBA SE rep today. :)
 
You guys should do a blog post about backups for SQL Server. Explain Full vs Simple recovery models, then (maybe a separate post) explain how to setup regular scheduled backups correctly when using the Full recovery model. I wasn't having much luck figuring it out on my own until I came here
 
@MikeFal marian gave him the gun
@Rachel well those are quite plentiful. the real problem is when people think that backup -> truncate is an appropriate methodology to maintain storage
 
7:41 PM
I would like to say to this user. Well you have a denormalized table, what do you expect
Fine but when I have 1000 rows in my table I should using select 1000x5 soo my table will slow — user2011304 2 hours ago
 
I was all ready to defend table DE-normalization until I saw the table....
 
@bluefeet done
 
Yay, time to go work from home and prep for a long night of watching SP2 install.
 
That's not the sort of denormalizing I would do.
 
@MikeFal cool runnings
 
7:52 PM
0
Q: SQL Injection to create user, Tried lots

SlimmonsI have a site that one of my friends was practicing SQL Injections on, and they were able to pull all of my dummy database information out (which I was able to do as well), but I was unable to create a new user. I'm only placing the SQL Injections into the username/password fields, and not into ...

seems legit?
 
@Lamak meh
@Lamak have you upvoted @bluefeet yet?
2
A: JOIN 2 Tables and SELECT more cell

bluefeetYour person table is not normalized so in order to retrieve data you will have to join multiple times like you found or you can use a UNION ALL query to unpivot the data. This type of query will be similar to this: select * from work w inner join ( select person1 p_name from person union ...

 
@swasheck now I have
 
00:00 - 20:0020:00 - 00:00

« first day (752 days earlier)      last day (4127 days later) »