@MikeFal sorry missed this. I only know of things like NHL GameCenter Live and NBC Sports Network has something similar. Not sure how much they cost beyond minimal access, and not sure they work where you are (like Hulu, Netflix etc). I've come to rely on my Slingbox when traveling - works very well.
@JNK sorry about that. Fooled me too. We're going to fix it.
It can be on a monitored instance, but if you are collecting a high number of databases / database files, have low thresholds for top SQL, or have a ton of indexes and have frag man turned on, it can influence the status of that server.
Size of your database doesn't matter, it's about how much data you're collecting - influenced by a variety of things (mainly the factors I just mentioned).
You should expect to collect about 2 GB long term for each server you monitor, but that can bend either way, again depending. The space used ramps up quickly but plateaus, because we start aggregating and discarding granularity over time. You can customize that rate.
You'll need to set up the monitoring service to run with a user that has access to both (think domain trusts). Most people who have issues here run the monitoring service in the domain where the servers run, and have a tunnel for it to communicate to the database in their DMZ or office.
Or they just put the database in the data center too.
@JNK the service itself has to run as a Windows account (domain or workstation). You can connect to SQL Server using SQL auth, but the service still needs to connect to the OS for some of the data (that's where the Windows account comes in).
It's all a trade-off. Also you can remind him (since he should know), there is nothing to "put on" the production servers - it's agentless and only requires access, nothing gets installed on monitored servers.
@PaulWhite its fine, I think the powers that be are starting to see it too
He argued HARD about some of this failover stuff in a meeting that all the managers were in, and the CIO made a point yesterday of verifying that it actually worked correctly
@AaronBertrand can you link me to your article about the overhead of SS?
@JNK we have 20-ish licenses for it. a few notable instances: we run it on our investment data warehouse (10TB DSS system) and the instance that backs our real-time trading platform (you know, the one where milliseconds matter?), and our instance that serves our college investment program (customer-facing, real time demand). the extremely negligible overhead from performance advisor is worth what it's provided for us.
you guys should be a team and, from what i can tell, you're not an ignorant moron (quite the opposite) so i'm not sure why this guy poohpoohs so many things.
but cheer up. he'll retire soon, right? or be retired?
@JNK i tend to not talk to my wife about it (mostly due to the fact that i dont like talking to people about things that they're not interested in)... i just come home and have a double and she knows: bad day. :)
I don't even know where to start, to be honest. It's like he's trying to get to San Francisco on a tricycle, and I passed him on my bike, and I'm going in the other direction.
@PaulWhite thanks, appreciated. I don't think this is a plan problem though, deep down. This is just terribly inefficient query and/or table design.
@PaulWhite In my mind I always equate questions/problems like that one with "Why won't this car that my 10 year old built in their yard do 70 mph like other cars?"
I have a lengthy T-SQL Script. When SSMS attempts to parse the script, it raises this error:
Msg 565, Level 18, State 1, Line 106620
A stack overflow occurred in the server while compiling the query. Please simplify the query.
The only thing I can see to simplify my query is to remove portion...
@JNK wow. the problem is that you're arguing from a position of intelligence but not influence. hopefully all of this is eroding his influence and increasing yours.
@AaronBertrand Yes, the returned message comes from SQL Server not SSMS. Big string concatenations and huge UNION ALL clauses are about the only two ways I know of to get an error from parsing/compilation.
@PaulWhite I had fun with this when I was migrating MySQL. Because MySQL's "backups" - at least in the WordPress world - are text files full of thousands of INSERT INTO statements. Fun times.
You can change logging to verbose in SQLSentryServer.exe.config but that should be reserved for being really stuck. Support will be a better resource than me if you get to that point.
The client only needs to be able to access the SQL Sentry database, not the rest of it, and you can install as many clients on as many machines as you like. The only time you need to access the monitored server is if you do something like kill some session or try to change a job or what have you
View A uses table 1. View B also uses T1. View C uses Views A and B and finally View D uses C. The data type in T1 changed from char to varchar and metadata in View D is stale.
Does sys.sp_refreshsqlmodule against View D cascade to all views that comprise it or does my friend need to run this per module? Documentation doesn't specify
which of our monitored "Top SQL" queries were running during an overnight batch cycle. some consultant in india fired off a cognos query from hell and it collided with the overnight cycle. blame the dba. dba holds up a mirror and says, "look more closely"
@JNK I'm not crazy about coding style, much of it pre-dates me, and I've only been involved in improvement of a couple of stored procedures. A very large portion is "works fine, leave it alone" and I'm ok with that.
@JNK if you really want your mind blown, it will be much more efficient to get a one hour demo from Sales Engineering. They will point you to all the important things and explain why/how etc. Stumbling upon things is fun, but...
@JNK I promise you Scott is not a salesperson. I'm not even sure why we call the department "sales engineering" - he's technical and gives you the no-bullshit, technical run-down.
For two servers though, really, it's going to be what's listed on the site. He'll probably try to get you into a 5-pack, because it's basically five for the price of three.
Not like pressure you into it, just offer that as one of the options.
And even if you don't need to monitor them all the time, what some customers do is have an extra license or two that can float between different servers for different things.
Ugh, I hate PDO / PHP / MySQL. Trying to optimize this junk. Why does PDO send this as prepared and cache a plan for each parameter? Stupid stupid stupid...
Ok. Somebody help me out here. I can't figure out the difference between database-design and database-schema. Can someone explain the difference to me? I could possibly see database schema referring to the "physical" object in a database, but the wiki doesn't seem to define it that way.