I have worked on SQL stored procedures and I have noticed that many people use two different approaches -
First, to use select queries i.e. something like
Select * from TableA where colA = 10 order by colA
Second, is to do the same by constructing a query i.e. like
Declare #sqlstring varchar(100)
Declare #sqlwhereclause varchar(100)
Declare #sqlorderby varchar(100)
Set #sqlstring = 'Select * from TableA '
Set #sqlwhereclause = 'where colA = 10 '
Set #sqlorderby = 'order by colA'
Set #sqlstring = #sqlstring + #sqlwhereclause + #sqlorderby
exec #sqlstring
Now, I know both work fine. But, the second method I mentioned is a little annoying to maintain.
I want to know which one is better? Is there any specific reason one would resort to one method over the other? Any benefits of one method over other?
Use the first one. This will allow a query plan to be cached properly, apart from being the way you are supposed to work with SQL.
The second one is open to SQL Injection attacks, apart from the other issues.
With the dynamic SQL you will not get compile time checking, so it may fail only when invoked (the sooner you know about incorrect syntax, the better).
And, you noted yourself, the maintenance burden is also higher.
The second method has the obvious drawback of not being syntax checked at compile time. It does however allow a dynamic order by clause, which the first does not. I recommend that you always use the first example unless you have a very good reason to make the query dynamic. And, as #Oded has already pointed out, be sure to guard yourself against sql injection if you do go for the second approach.
I don't have a full comprehensive answer for you, but I can tell you right now that the latter method is much more difficult to work with when importing the stored procedure as a function in an ORM. Since the SQL is constructed dynamically, you have to manually create any type-classes that are returned from the stored procedure that aren't directly correlated to entities in your model.
With that in mind, there are times where you simply can't avoid constructing a SQL statement, especially when where clauses and joins depend on the parameters passed in. In my experience, I have found that stored procs that are creating large, variably joined/whered statements for EXECs are trying to do too many things. In these situations, I would recommend you keep the Single Responsibility Principle in mind.
Executing dynamic SQL inside a stored procedure reduces the value of using stored procedures to just a saved query container. Stored procedures are mostly beneficial in that the query execution plan (a very costly operation) is compiled and stored in memory the first time the procedure is executed. This means that every subsequent execution of the procedure is bypassing the query plan calculations, and jumping right to the data retrieval portiion of the operation.
Also, allowing a stored procedure to take an executable query string as a parameter is dangerous. Anyone with execute permission on granted on the procedure could potentially cause havoc on the rest of the database.
Related
I want to insert the results of a stored procedure into a temp table using OPENROWSET. However, the issue I run into is I'm not able to pass parameters to my stored procedure.
This is my stored procedure:
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
ALTER PROCEDURE [dbo].[N_spRetrieveStatement]
#PeopleCodeId nvarchar(10),
#StatementNumber int
AS
SET NOCOUNT ON
DECLARE #PersonId int
SELECT #PersonId = [dbo].[fnGetPersonId](#PeopleCodeId)
SELECT *
INTO #tempSpRetrieveStatement
FROM OPENROWSET('SQLNCLI', 'Server=PCPRODDB01;Trusted_Connection=yes;',
'EXEC Campus.dbo.spRetrieveStatement #StatementNumber, #PersonId');
--2577, 15084
SELECT *
FROM #tempSpRetrieveStatement;
OpenRowSet will not allow you to execute Procedure with input parameters. You have to use INSERT/EXEC.
INTO #tempSpRetrieveStatement(Col1, Col2,...)
EXEC PCPRODDB01.Campus.dbo.spRetrieveStatement #StatementNumber, #PersonId
Create and test a LinkedServer for PCPRODDB01 before running the above command.
The root of your problem is that you don't actually have parameters inside your statement that you're transmitting to the remote server you're connecting to, given the code sample you provided. Even if it was the very same machine you were connecting to, they'd be in different processes, and the other process doesn't have access to your session variables.
LinkedServer was mentioned as an option, and my understanding is that's the preferred option. However in practice that's not always available due to local quirks in tech or organizational constraints. It happens.
But there is a way to do this.
It's hiding in plain sight.
You need to pass literals into the string that will be executed on the other server, right?
So, you start by building the string that will do that.
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
ALTER PROCEDURE [dbo].[N_spRetrieveStatement]
#PeopleCodeId nvarchar(10),
#StatementNumber int
AS
SET NOCOUNT ON
DECLARE
#PersonId INT,
#TempSQL VARCHAR(4000) = '';
SELECT #PersonId = [dbo].[fnGetPersonId](#PeopleCodeId);
SET #TempSQL =
'EXEC Campus.dbo.spRetrieveStatement(''''' +
FORMAT(#StatementNumber,'D') +''''', ''''' +
FORMAT(#PersonId,'D') + ''''')';
--2577, 15084
Note the seemingly excessive number of quotes. That's not a mistake -- that's foreshadowing. Because, yes, OPENROWSET hates taking variables as parameters. It, too, only wants literals. So, how do we give OPENROWSET what it needs?
We create a string that is the entire statement, no variables of any kind. And we execute that.
SET #TempSQL =
'SELECT * INTO #tempSpRetrieveStatement ' +
'FROM OPENROWSET(''SQLNCLI'', ''Server=PCPRODDB01;Trusted_Connection=yes;'', ' + #TempSQL +
'EXEC Campus.dbo.spRetrieveStatement #StatementNumber, #PersonId';
EXEC (#TempSQL);
SELECT *
FROM #tempSpRetrieveStatement;
And that's it! Pretty simple except for counting your escaped quotes, right?
Now... This is almost beyond the scope of the question you asked, but it is a 'gotcha' I've experienced in executing stored procedures in another machine via OPENROWSET. You're obviously used to using temp tables. This will fail if the stored procedure you're calling is creating temp tables or doing a few other things that -- in a nutshell -- inspire the terror of ambiguity into your SQL server. It doesn't like ambiguity. If that's the case, you'll see a message like this:
"Msg 11514, Level 16, State 1, Procedure sp_describe_first_result_set, Line 1
The metadata could not be determined because statement '…your remote EXEC statement here…' in procedure '…name of your local stored procedure here…' contains dynamic SQL. Consider using the WITH RESULT SETS clause to explicitly describe the result set."
So, what's up with that?
You don't just get data back with OPENROWSET. The local and remote servers have a short conversation about what exactly the local server is going to expect from the remote server (so it can optimize receiving and processing it as it comes in -- something that's extremely important for large rowsets). Starting with SQL Server 2012, sp_describe_first_result_set is the newer procedure for this, and normally it executes quickly without you noticing it. It's just that it's powers of divination aren't unlimited. Namely, it doesn't know how to get the type and name information regarding temp tables (and probably a few other things it can't do -- PIVOT in a select statement is probably right out).
I specifically wanted to be sure to point this out because of your reply regarding your hesitation about using LinkedServer. In fact, the very same reasons you're hesitant are likely to render that error message's suggestion completely useless -- you can't even predict what columns you're getting and in what order until you've got them.
I think what you're doing will work if, say, you're just branching upstream based on conditional statements and are executing one of several potential SELECT statements. I think it will work if you're just not confident that you can depend on the upstream component being fixed and are trying to ensure that even if it varies, this procedure doesn't have to because it's very generic.
But on the other hand you're facing a situation in which you literally cannot guarantee that SQL Server can predict the columns, you're likely going to have to force some changes in the stored procedure you're calling to insist that it's stable. You might, for instance work out how to ensure all possible fields are always present by using CASE expressions rather than any PIVOT. You might create a session table that's dedicated to housing what you need to SELECT just long enough to do that then DELETE the contents back out of there. You might change the way in which you transmit your data such that it's basically gone through the equivalent of UNPIVOT. And after all that extra work, maybe it'll be just a matter of preference if you use LinkedServer or OPENROWSET to port the data across.
So that's the answer to the literal question you asked, and one of the limits on what you can do with the answer.
I am writing a sql procedure where almost everything will be dynamic including selecting, grouping, ordering by, and where clauses using IN statements. In terms of code reuse, readability, and maintenance it makes a lot of sense to just pass in an sql query as a string and execute it. I am writing my procedure right now so that all the relevant data is joined and formatted in a static query and then inserted into a table variable. I then want to pass in sql queries to be executed against the table variable.
This opens me up to sql injection in a big way. I could create table value parameters for each of the many parameter types I am passing in but I don't want to do that. What I would really like to be able to do sandbox my procedure in a such a way that, on the procedure level, it is only possible to do these things I want to allow; ie select from certain tables, but not grant permissions or anything funny like that. Can this be done?
Of course it can be done. It's a simple matter of programming. You would keep rules in tables, and write logic in your stored procedure to query the rules tables, and follow the rules.
It will be a monumental job that will basically amount to you writing custom code to do what SQL Server already does for you if you don't use a generic, dynamic stored procedure.
I wouldn't do it, but you don't have to let that stop you.
If yes, why are there still so many successful SQL injections? Just because some developers are too dumb to use parameterized statements?
When articles talk about parameterized queries stopping SQL attacks they don't really explain why, it's often a case of "It does, so don't ask why" -- possibly because they don't know themselves. A sure sign of a bad educator is one that can't admit they don't know something. But I digress.
When I say I found it totally understandable to be confused is simple. Imagine a dynamic SQL query
sqlQuery='SELECT * FROM custTable WHERE User=' + Username + ' AND Pass=' + password
so a simple sql injection would be just to put the Username in as ' OR 1=1--
This would effectively make the sql query:
sqlQuery='SELECT * FROM custTable WHERE User='' OR 1=1-- ' AND PASS=' + password
This says select all customers where they're username is blank ('') or 1=1, which is a boolean, equating to true. Then it uses -- to comment out the rest of the query. So this will just print out all the customer table, or do whatever you want with it, if logging in, it will log in with the first user's privileges, which can often be the administrator.
Now parameterized queries do it differently, with code like:
sqlQuery='SELECT * FROM custTable WHERE User=? AND Pass=?'
parameters.add("User", username)
parameters.add("Pass", password)
where username and password are variables pointing to the associated inputted username and password
Now at this point, you may be thinking, this doesn't change anything at all. Surely you could still just put into the username field something like Nobody OR 1=1'--, effectively making the query:
sqlQuery='SELECT * FROM custTable WHERE User=Nobody OR 1=1'-- AND Pass=?'
And this would seem like a valid argument. But, you would be wrong.
The way parameterized queries work, is that the sqlQuery is sent as a query, and the database knows exactly what this query will do, and only then will it insert the username and passwords merely as values. This means they cannot effect the query, because the database already knows what the query will do. So in this case it would look for a username of "Nobody OR 1=1'--" and a blank password, which should come up false.
This isn't a complete solution though, and input validation will still need to be done, since this won't effect other problems, such as XSS attacks, as you could still put javascript into the database. Then if this is read out onto a page, it would display it as normal javascript, depending on any output validation. So really the best thing to do is still use input validation, but using parameterized queries or stored procedures to stop any SQL attacks.
The links that I have posted in my comments to the question explain the problem very well. I've summarised my feelings on why the problem persists, below:
Those just starting out may have no awareness of SQL injection.
Some are aware of SQL injection, but think that escaping is the (only?) solution. If you do a quick Google search for php mysql query, the first page that appears is the mysql_query page, on which there is an example that shows interpolating escaped user input into a query. There's no mention (at least not that I can see) of using prepared statements instead. As others have said, there are so many tutorials out there that use parameter interpolation, that it's not really surprising how often it is still used.
A lack of understanding of how parameterized statements work. Some think that it is just a fancy means of escaping values.
Others are aware of parameterized statements, but don't use them because they have heard that they are too slow. I suspect that many people have heard how incredibly slow paramterized statements are, but have not actually done any testing of their own. As Bill Karwin pointed out in his talk, the difference in performance should rarely be used as a factor when considering the use of prepared statements. The benefits of prepare once, execute many, often appear to be forgotten, as do the improvements in security and code maintainability.
Some use parameterized statements everywhere, but with interpolation of unchecked values such as table and columns names, keywords and conditional operators. Dynamic searches, such as those that allow users to specify a number of different search fields, comparison conditions and sort order, are prime examples of this.
False sense of security when using an ORM. ORMs still allow interpolation of SQL statement parts - see 5.
Programming is a big and complex subject, database management is a big and complex subject, security is a big and complex subject. Developing a secure database application is not easy - even experienced developers can get caught out.
Many of the answers on stackoverflow don't help. When people write questions that use dynamic SQL and parameter interpolation, there is often a lack of responses that suggest using parameterized statements instead. On a few occasions, I've had people rebut my suggestion to use prepared statements - usually because of the perceived unacceptable performance overhead. I seriously doubt that those asking most of these questions are in a position where the extra few milliseconds taken to prepare a parameterized statement will have a catastrophic effect on their application.
Well good question.
The answer is more stochastic than deterministic and I will try to explain my view, using a small example.
There many references on the net that suggest us to use parameters in our queries or to use stored procedure with parameters in order to avoid SQL Injection (SQLi). I will show you that stored procedures (for instance) is not the magic stick against SQLi. The responsibility still remains on the programmer.
Consider the following SQL Server Stored Procedure that will get the user row from a table 'Users':
create procedure getUser
#name varchar(20)
,#pass varchar(20)
as
declare #sql as nvarchar(512)
set #sql = 'select usrID, usrUName, usrFullName, usrRoleID '+
'from Users '+
'where usrUName = '''+#name+''' and usrPass = '''+#pass+''''
execute(#sql)
You can get the results by passing as parameters the username and the password. Supposing the password is in free text (just for simplicity of this example) a normal call would be:
DECLARE #RC int
DECLARE #name varchar(20)
DECLARE #pass varchar(20)
EXECUTE #RC = [dbo].[getUser]
#name = 'admin'
,#pass = '!#Th1siSTheP#ssw0rd!!'
GO
But here we have a bad programming technique used by the programmer inside the stored procedure, so an attacker can execute the following:
DECLARE #RC int
DECLARE #name varchar(20)
DECLARE #pass varchar(20)
EXECUTE #RC = [TestDB].[dbo].[getUser]
#name = 'admin'
,#pass = 'any'' OR 1=1 --'
GO
The above parameters will be passed as arguments to the stored procedure and the SQL command that finally will be executed is:
select usrID, usrUName, usrFullName, usrRoleID
from Users
where usrUName = 'admin' and usrPass = 'any' OR 1=1 --'
..which will get all rows back from users
The problem here is that even we follow the principle "Create a stored procedure and pass the fields to search as parameters" the SQLi is still performed. This is because we just copy our bad programming practice inside the stored procedure. The solution to the problem is to rewrite our Stored Procedure as follows:
alter procedure getUser
#name varchar(20)
,#pass varchar(20)
as
select usrID, usrUName, usrFullName, usrRoleID
from Users
where usrUName = #name and usrPass = #pass
What I am trying to say is that the developers must learn first what an SQLi attack is and how can be performed and then to safeguard their code accordingly. Blindly following 'best practices' is not always the safer way... and maybe this is why we have so many 'best practices'- failures!
Yes, the use of prepared statements stops all SQL injections, at least in theory. In practice, parameterized statements may not be real prepared statements, e.g. PDO in PHP emulates them by default so it's open to an edge case attack.
If you're using real prepared statements, everything is safe. Well, at least as long as you don't concatenate unsafe SQL into your query as reaction to not being able to prepare table names for example.
If yes, why are there still so many successful SQL injections? Just because some developers are too dumb to use parameterized statements?
Yes, education is the main point here, and legacy code bases. Many tutorials use escaping and those can't be easily removed from the web, unfortunately.
I avoid absolutes in programming; there is always an exception. I highly recommend stored procedures and command objects. A majority of my back ground is with SQL Server, but I do play with MySql from time to time. There are many advantages to stored procedures including cached query plans; yes, this can be accomplished with parameters and inline SQL, but that opens up more possibilities for injection attacks and doesn't help with separation of concerns. For me it's also much easier to secure a database as my applications generally only have execute permission for said stored procedures. Without direct table/view access it's much more difficult to inject anything. If the applications user is compromised one only has permission to execute exactly what was pre-defined.
My two cents.
I wouldn't say "dumb".
I think the tutorials are the problem. Most SQL tutorials, books, whatever explain SQL with inlined values, not mentioning bind parameters at all. People learning from these tutorials don't have a chance to learn it right.
Because most code isn't written with security in mind, and management, given a choice between adding features (especially something visible that can be sold) and security/stability/reliability (which is a much harder sell) they will almost invariably choose the former. Security is only a concern when it becomes a problem.
Can parameterized statement stop all SQL injection?
Yes, as long as your database driver offers a placeholder for the every possible SQL literal. Most prepared statement drivers don't. Say, you'd never find a placeholder for a field name or for an array of values. Which will make a developer to fall back into tailoring a query by hand, using concatenation and manual formatting. With predicted outcome.
That's why I made my Mysql wrapper for PHP that supports most of literals that can be added to the query dynamically, including arrays and identifiers.
If yes, why are there still so many successful SQL injections? Just because some developers are too dumb to use parameterized statements?
As you can see, in reality it's just impossible to have all your queries parameterized, even if you're not dumb.
First my answer to your first question: Yes, as far as I know, by using parameterized queries, SQL injections will not be possible anymore. As to your following questions, I am not sure and can only give you my opinion on the reasons:
I think it's easier to "just" write the SQL query string by concatenate some different parts (maybe even dependent on some logical checks) together with the values to be inserted.
It's just creating the query and executing it.
Another advantage is that you can print (echo, output or whatever) the sql query string and then use this string for a manual query to the database engine.
When working with prepared statements, you always have at least one step more:
You have to build your query (including the parameters, of course)
You have to prepare the query on the server
You have to bind the parameters to the actual values you want to use for your query
You have to execute the query.
That's somewhat more work (and not so straightforward to program) especially for some "quick and dirty" jobs which often prove to be very long-lived...
Best regards,
Box
SQL injection is a subset of the larger problem of code injection, where data and code are provided over the same channel and data is mistaken for code. Parameterized queries prevent this from occurring by forming the query using context about what is data and what is code.
In some specific cases, this is not sufficient. In many DBMSes, it's possible to dynamically execute SQL with stored procedures, introducing a SQL injection flaw at the DBMS level. Calling such a stored procedure using parameterized queries will not prevent the SQL injection in the procedure from being exploited. Another example can be seen in this blog post.
More commonly, developers use the functionality incorrectly. Commonly the code looks something like this when done correctly:
db.parameterize_query("select foo from bar where baz = '?'", user_input)
Some developers will concatenate strings together and then use a parameterized query, which doesn't actually make the aforementioned data/code distinction that provides the security guarantees we're looking for:
db.parameterize_query("select foo from bar where baz = '" + user_input + "'")
Correct usage of parameterized queries provides very strong, but not impenetrable, protection against SQL injection attacks.
To protect your application from SQL injection, perform the following steps:
Step 1. Constrain input.
Step 2. Use parameters with stored procedures.
Step 3. Use parameters with dynamic SQL.
Refer to http://msdn.microsoft.com/en-us/library/ff648339.aspx
even if
prepared statements are properly used throughout the web application’s own
code, SQL injection flaws may still exist if database code components construct
queries from user input in an unsafe manner.
The following is an example of a stored procedure that is vulnerable to SQL
injection in the #name parameter:
CREATE PROCEDURE show_current_orders
(#name varchar(400) = NULL)
AS
DECLARE #sql nvarchar(4000)
SELECT #sql = ‘SELECT id_num, searchstring FROM searchorders WHERE ‘ +
‘searchstring = ‘’’ + #name + ‘’’’;
EXEC (#sql)
GO
Even if the application passes the user-supplied name value to the stored
procedure in a safe manner, the procedure itself concatenates this directly into
a dynamic query and therefore is vulnerable.
I am using SQL Server 2008 Enterprise on Windows Server 2008 Enterprise. In a stored procedure, we can execute a SELECT statement directly. And it could also be executed in this new way, I am wondering which method is better, and why?
New method,
declare #teststatement varchar(500)
set #teststatement = 'SELECT * from sometable'
print #teststatement
exec (#teststatement)
Traditional method,
SELECT * from sometable
regards,
George
FYI: it’s not a new method, it is known as Dynamic SQL.
Dynamic SQL are preferred when we need to set or concatenate certain values into sql statements.
Traditional or normal way sql statements are recommended, because stored procedures are complied. Complied on first run "Stored Procedure are Compiled on First Run"
, execution plan of statements are being created at the time of compilation.
Dynamic sqls are ignored while creating execution plans, because it is taken as string (VARCHAR or NVARCHAR as declared).
Refer following articles for more details about dynamic query and stored procs
Introduction to Dynamic SQL Part 1
Introduction to Dynamic SQL Part 2
Everything you wanted to know about Stored Procedures
The traditional method is safer, because the query is parsed when you save it. The query in the 'exec' method is not parsed and can contain errors.
The "new" way, as mentioned, has nothing to do with SQL 2008. EXEC has been available for quite some time. It's also - in most cases - a Very Bad Idea.
You lose parameterization - meaning you are now vulnerable to SQL Injection. It's ugly and error-prone. It's less efficient. And it creates a new execution scope - meaning it can't share variables, temp tables, etc. - from it's calling stored proc.
sp_executesql is another (and preferred) method of executing dynamic SQL. It's what your client apps use, and it supports parameters - which fixes the most glaring problem of EXEC. However, it too has very limited use cases within a stored proc. About the only redeeming use is when you need a dynamic table or column name. T-SQL does not support a variable for that - so you need to use sp_executesql. The number of times you need or should be doing that are very low.
Bottom line - you'd be best off forgetting you ever heard of it.
I know this has something to do with parameter sniffing, but I'm just perplexed at how something like the following example is even possible with a piece of technology that does so many complex things well.
Many of us have run into stored procedures that intermittently run several of orders of magnitude slower than usual, and then if you copy out the sql from the procedure and use the same parameter values in a separate query window, it runs as fast as usual.
I just fixed a procedure like that by converting this:
alter procedure p_MyProc
(
#param1 int
) as -- do a complex query with #param1
to this:
alter procedure p_MyProc
(
#param1 int
)
as
declare #param1Copy int;
set #param1Copy = #param1;
-- Do the query using #param1Copy
It went from running in over a minute back down to under one second, like it usually runs. This behavior seems totally random. For 9 out of 10 #param1 inputs, the query is fast, regardless of how much data it ends up needing to crunch, or how big the result set it. But for that 1 out of 10, it just gets lost. And the fix is to replace an int with the same int in the query?
It makes no sense.
[Edit]
#gbn linked to this question, which details a similar problem:
Known issue?: SQL Server 2005 stored procedure fails to complete with a parameter
I hesitate to cry "Bug!" because that's so often a cop-out, but this really does seem like a bug to me. When I run the two versions of my stored procedure with the same input, I see identical query plans. The only difference is that the original takes more than a minute to run, and the version with the goofy parameter copying runs instantly.
The 1 in 10 gives the wrong plan that is cached.
RECOMPILE adds an overhead, masking allows each parameter to be evaluated on it's own merits (very simply).
By wrong plan, what if the 1 in 10 generates an scan on index 1 but the other 9 produce a seek on index 2? eg, the 1 in 10 is, say, 50% of the rows?
Edit: other questions
Known issue?: SQL Server 2005 stored procedure fails to complete with a parameter
Stored Procedure failing on a specific user
Edit 2:
Recompile does not work because the parameters are sniffed at compile time.
From other links (pasted in):
This article explains...
...parameter values are sniffed during compilation or recompilation...
Finally (edit 3):
Parameter sniffing was probably a good idea at the time and probably works well mostly. We use it across the board for any parameter that will end up in a WHERE clause.
We don't need to use it because we know that only a few (more complex eg reports or many parameters) could cause issues but we use it for consistency.
And the fact that it will come back and bite us when the users complain and we should have used masking...
It's probably caused by the fact that SQL Server compiles stored procedures and caches execution plans for them and the cached execution plan is probably unsuitable for this new set of parameters. You can try WITH RECOMPILE option to see if it's the cause.
EXECUTE MyProcedure [parameters] WITH RECOMPILE
WITH RECOMPILE option will force SQL Server to ignore the cached plan.
I have had this problem repeatedly on moving my code from a test server to production - on two different builds of SQL Server 2005. I think there are some big problems with the parameter sniffing in some builds of SQL Server 2005. I never had this problem on the dev server, or on two local developer edition boxes. I've never seen it it be such a big problem on SQL Server 2000 or any version going back to 6.5 either.
The cases where I found it, the only workaround was to use parameter masking, and I'm still hoping the DBAs will patch up the production server to SP3 so it will maybe go away. Things which did not work:
using the WITH RECOMPILE hint on EXEC or in the SP itself.
dropping and recreating the SP
using sp_recompile
Note that in the case I was working on, the data was not changing since an earlier invocation - I had simply scripted the code onto the production box which already had data loaded. All the invocations came with no changes to the data since before the SPs existed.
Oh, and if SQL Server can't handle this without masking, they need to add a parameter modifier NOSNIFF or something. What happens if you mask all your parameters, so you have #Something_parm and #Something_var and someone changes the code to use the wrong one and all of a sudden you have a sniffing problem again? Plus you are polluting the namespace within the SP. All these SPs I am "fixing" drive me nuts because I know they are going to be a maintenance nightmare for the less experienced satff I will be handing this project off to one day.
Could you check on the SQL Profiler how many reads and execution time when it is quick and when it is slow? It could be related to the number of rows fetched depending on the parameter value. It doesn't sound like a cache plan issue.
I know this is a 2 year old thread, but it might help someone down the line.
Once you analyze the query execution plans and know what the difference is between the two plans (query by itself and query executing in the stored procedure with a flawed plan), you can modify the query within the stored procedure with a query hint to resolve the issue. This works in a scenario where the query is using the incorrect index when executed in the stored procedure. You would add the following after the table in the appropriate location of your procedure:
SELECT col1, col2, col3
FROM YourTableHere WITH (INDEX (PK_YourIndexHere))
This will force the query plan to use the correct index which should resolve the issue. This does not answer why it happens but it does provide a means to resolve the issue without worrying about copying the parameters to avoid parameter sniffing.
As indicated it be a compilation issue. Does this issue still occur if you revert the procedure? One thing you can try if this occurs again to force a recompilation is to use:
sp_recompile [ #objname = ] 'object'
Right from BOL in regards to #objname parameter:
Is the qualified or unqualified name of a stored procedure, trigger, table, or view in the current database. object is nvarchar(776), with no default. If object is the name of a stored procedure or trigger, the stored procedure or trigger will be recompiled the next time that it is run. If object is the name of a table or view, all the stored procedures that reference the table or view will be recompiled the next time they are run.
If you drop and recreate the procedure you could cause clients to fail if they try and execute the procedure. You will also need to reapply security settings.
Is there any chance that the parameter value being provided is sometimes not int?
Is every query reference to the parameter comparing it with int values, without functions and without casting?
Can you increase the specificity of any expressions using the parameter to make the use of multifield indexes more likely?
It is a problem with plan caching, and it isn't always related to parameters, as it was in your scenario.
(Parameter Sniffing problems occur when a proc is called with unusual parameters the FIRST time it runs, and so the cached plan works great for those odd values, but lousy for most other times the proc is called.)
We had a similar situation when the app team deleted all old records from a highly-used log table on a production server. Removing records improves performance, right? Nope, performance immediately tanked.
Turns out that a frequently-used stored proc was recompiled right when the table was nearly empty, and it cached an extremely poor execution plan ("hey, there's only 50 records here, might as well do a Table Scan!"). Would have happened no matter what the initial parameters.
Our fix was to force a recompile with sp_recompile.