Tuesday, January 08, 2019

Print.. the disruptor of batch deletes in SQL

Someone had an issue where a batched delete script was not deleting anything. I looked over some code in our repository and noticed two patterns the way queries are written to handle batch deletes

One is a while loop that runs while @@rowcount is greater than 0

WHILE @@rowcount > 0
  DELETE TOP (5000)
  FROM SomeTable

The other way is to run a while loop which is always true and then check if @@rowcount is 0, if it is 0 then break out of the loop

 WHILE 1 = 1  
        DELETE TOP(5000)  
        FROM SomeTable

        IF @@ROWCOUNT = 0  

I have always used WHILE @@rowcount > 0 but you have to be careful because @@rowcount could be 0 when your while loop starts

Let's take a look at an example. This is a simplified example without a where clause..but let's say you have to delete several million rows from a table with many more millions of rows and the table is replicated... in that case you want to batch the deletes so that your log file doesn't fill up, replication has a chance to catch up and in general the deletes should run faster

SELECT TOP 20000 row_number() OVER(ORDER BY t1.id) AS SomeId,
  getutcdate() AS SomeDate, newid() AS SomeValue
INTO SomeTable
FROM sys.sysobjects t1
CROSS JOIN sys.sysobjects t2

SELECT COUNT(*) from SomeTable 

SELECT * FROM SomeTable WHERE 1= 0

WHILE @@rowcount > 0
  DELETE TOP (5000)
  FROM SomeTable

SELECT COUNT(*) from SomeTable 

DROP TABLE SomeTable -- Added here as cleanup in case people run the example

This is of course a silly example because why would you do a count like that against a different table before a delete

But what if you had this instead, you put a nice print statement there so that from the output you see when it started and you would also see the rowcounts?

SELECT TOP 20000 row_number() OVER(ORDER BY t1.id) AS SomeId,
  getutcdate() AS SomeDate, newid() AS SomeValue
INTO SomeTable
FROM sys.sysobjects t1
CROSS JOIN sys.sysobjects t2

SELECT COUNT(*) from SomeTable 

PRINT' Starting my update now....'

WHILE @@rowcount > 0
  DELETE TOP (5000)
  FROM SomeTable

SELECT COUNT(*) from SomeTable 

DROP TABLE SomeTable -- Added here as cleanup in case people run the example

The count is 20000 before and after the loop, nothing got delete, this is because a print statement will reset @@rowcount  to 0

Take a look by running this simple set of queries

SELECT @@rowcount as 'Rowcount'
SELECT @@rowcount as 'RowcountAfterPrint'

Here is what the output looks like

After the  PRINT line @@rowcount is reset back to 0

So if you want to use a while loop while checking @@rowcount, do this instead by running the delete first once outside the loop

SELECT TOP 20000 row_number() OVER(ORDER BY t1.id) AS SomeId,
  getutcdate() AS SomeDate, newid() AS SomeValue
INTO SomeTable
FROM sys.sysobjects t1
CROSS JOIN sys.sysobjects t2

SELECT COUNT(*) from SomeTable 

PRINT' Starting my update now....'

FROM SomeTable

WHILE @@rowcount > 0
  DELETE TOP (5000)
  FROM SomeTable

SELECT COUNT(*) from SomeTable 

DROP TABLE SomeTable -- Added here as cleanup in case people run the example

If you run the delete this way if there was something to delete, the while loop would be entered, if the table was empty then there would no need to enter the while loop

Also keep in mind that it is not just PRINT that will reset @@rowcount back to 0.

From Books On Line:


That's all... hopefully this helps someone out in the future if they notice nothing gets deleted

Friday, January 04, 2019

Proactive notifications

It has been a while since I wrote some of my best practices posts. I decided to revisit these posts again to see if anything has changed, I also wanted to see if I could add some additional info.

 Today we are going to look at proactive notifications

In the SQL Server Maintenance post from yesterday I touched upon proactive notifications a little, today I want to dive a little deeper into this subject.
The last thing you want to hear as a DBA is  ny of the following things from the end users or developers
  • The transaction log is full
  • The database is very slow
  • The latest backup we have is 9 days old
  • The table that was created has 2 extra columns this morning
  • Everything is locked up can't get any results back from a query
  • Deadlocks are occurring
What you really want to have at your shop is a tool like Quest FoglightSolarwindsRed Gate SQL Monitor or similar. The benefit of these tools is that there is a central location where you can look at all the alerts at a glance. You get a lot of stuff out of the box and all you have to do is tell it what server to start monitoring. I would suggest to start using the trial version to see if it is something that would be beneficial for your organization.
Of course you can roll your own solution as well, this will involve work and unless your time is worthless or you are bored out of your mind after work I wouldn't do it.

Utilize the logs

You need to scan the errorlog periodically to see if there are errors, you can automate this, no need to start opening log files every 5 minutes. create a SQL Agent job that runs every 5 minutes and checks if there are any errors since it last ran. You can use the xp_readerrorlog proc to read the error log from with sql server with T-SQL.
Here is a small example of what you can do if you have this in a SQL Agent job that runs every 5 minutes or so, you can of course email yourself the results, dump the result into a table that is perhaps shown on a dashboard in the office, there are many possibilities.

--This will hold the rows
CREATE TABLE #ErrorLog (LogDate datetime, ProcessInfo VarChar(10), 
ErrorMessage VarChar(Max))

-- Dump the errorlog into the table
EXEC master.dbo.xp_readerrorlog

-- Delete everything older than 5 minutes
-- ideally you will store the max date when it ran last
DELETE #ErrorLog

-- Some stuff you want to check for
-- Failed backups...you want to know this
WHERE ErrorMessage LIKE'BACKUP failed%'

-- Why does it take so looong to grow a file, maybe rethink your settings
WHERE ErrorMessage LIKE'Autogrow of file%'

-- What is going on any backups or statistic updates running at this time?
WHERE ErrorMessage LIKE'SQL Server has encountered %occurrence(s) of I/O requests taking longer than%'

-- My mirror might not be up to date
WHERE ErrorMessage LIKE'The alert for ''unsent log'' has been raised%'


Those are just small samples, you might want to look for other kind of messages from the errorlog

The transaction log is full

You want to make sure that you know you are running out of space before you run out of space. I covered this in the SQL Server Maintenance post Take a look at the sections Make sure that you have enough space left on the drives and Make sure that you have enough space left for the filegroups In those two section I described what to look for and also supplied code that you can then plug into your own solution

The database is very slow

This complaint you hear every now and then, I have seen this from time to time. There are several things that could be happening, here is a list
Someone decided to take a backup of that 1 TB database in the middle of the day
The update statistics job is still running
Statistics are stale and haven't been updated in a long time
The virus scan is running amok and nobody told it to ignore the database files
Someone decided to query all the data all at once
If you have a tool like Quest FoglightConfio IgniteRed Gate SQL Monitor or similar then you can see what query ran at what time, what it did and how long it ran.
You can of course also use sp_who2, BlkBy column and DBCC INPUTBUFFER to see what is going on
If you like to use Dynamic Management Views, then take a look at Glenn Berry's SQL Server 2005 Diagnostic Information Queries  posts, there is a .sql file in each post with all kind of queries to discover all kinds of stuff about your server.
It could also be that your hardware is having issues, make sure the IOs look good and check the eventlog for any clues.

The latest backup we have is 9 days old

The following query will give you for all the databases the last time it was backed up or display NEVER if it wasn't backed up

SELECT s.Name AS DatabaseName,'Database backup was taken on  ' + 
CASE WHEN MAX(b.backup_finish_date) IS NULL THEN 'NEVER!!!' ELSE
CONVERT(VARCHAR(12), (MAX(b.backup_finish_date)), 101) END AS LastBackUpTime
FROM sys.sysdatabases s
LEFT OUTER JOIN msdb.dbo.backupset b ON b.database_name = s.name

Here is what the output will look like

DatabaseName LastBackUpTime
--------------  ---------------------------------------
model         Database backup was taken on  NEVER!!!
msdb         Database backup was taken on  12/10/2012
ReportServer Database backup was taken on  NEVER!!!

As you can see that is not that great, all the databases should be backed up on a regular basis. Scroll up to the Utilize the logs section to see how you can check the errorlog for failed backup messages.

Everything is locked up, you can't get any results back from a query

Usually this indicates that there is an open transaction somewhere that has not finished or someone did the BEGIN TRAN part but never did a COMMIT or ROLLBACK.
Some people just restart the server to 'fix' the issue, of course if you do that you will never know what the root cause is and you never know when it will happen again.
We can easily show what happens when you have an open transaction, btw don't do this on the production server.

In 1 query window run this, replace SomeTable with a real table name.



You will get a message that the query completed successfully
In another window run this


That query won't return anything unless the first one is commited or rolled back

Now run this query below, the first column should have the text AWAITING COMMAND

SELECT   sys.cmd
FROM    sys.dm_tran_locks lok
JOIN    sys.dm_os_waiting_tasks wat
ON      lok.lock_owner_address = wat.resource_address
JOIN sys.sysprocesses sys ON wat.blocking_session_id = sys.spid

As you can see you have a blocking_session_id and a session_id, this will tell you which session_id is being blocked. You can now verify that the transaction session_id is blocking the other id
Go back to that first command window and execute a rollback


The query that had that second select should now be done as well, if you run that query that checks for the waits it should be clean as well.
Of course you could have done the same exercise by running sp_who2, looking at the BlkBy column, finding out what that session is doing by running DBCC INPUTBUFFER(session_id) with that session_id

Deadlocks are occurring

There is already a post written on LessThanDot explaining how you can get emailed when deadlocks occur. Ted Krueger wrote that post and it can be found here: Proactive Deadlock Notifications


I only touched the surface of what can be done in this post. I want you to find out if there is any monitoring being done in your shop, who gets notified? I have worked in places where the end user was the proactive notification, as long as we fixed it before the business users started to complaint life was good. Manual notifications and homebrew solutions might work for a while but when you add more and more servers and you add more people to the team this becomes laborious and error prone.

Thursday, January 03, 2019

Cursors and loops

It has been a while since I wrote some of my best practices posts. I decided to revisit these posts again to see if anything has changed, I also wanted to see if I could add some additional info.

 Today we are going to look at cursors

Why do we hate those poor cursors?

Let's first see why people tend to use cursors. Let's say you come from a procedural language and this is the first time you are using SQL. In the procedural language you know how to traverse a list, you of course will look for something that is similar in SQL........bingo!!! you found it...the almighty cursor....the crusher of almost all SQL Server performance. You start using it, your code works, you are happy, life is good.

Now a team member tells you that the cursor is evil and should never ever be used. You are confused, if a cursor is never to be used then why is it part of the language? Well you might say the same for the GOTO statement, this exists in SQL.  Edsger Dijkstra's letter Go To Statement Considered Harmful was published in the March 1968 Communications of the ACM.

The reason that cursors are evil is that they tend to be slower than a set based solution. Cursors are not needed for 99% of the cases. SQL is a set based language, it works best with sets of data, not row by row processing, when you do something set based it will generally perform hundreds of times faster than using a cursor.

Take a look at this code

CREATE FUNCTION dbo.cursorEnroll ()
        DECLARE @studentsEnrolled INT
        SET @studentsEnrolled = 0
        DECLARE myCursor CURSOR FOR
            SELECT enrollementID
                FROM courseEnrollment
        OPEN myCursor;
        FETCH NEXT FROM myCursor INTO @studentsEnrolled
        WHILE @@FETCH_STATUS = 0
                SET @studentsEnrolled = @studentsEnrolled+1
                    FETCH NEXT FROM myCursor INTO @studentsEnrolled
        CLOSE myCursor
        RETURN @studentsEnrolled

That whole flawed cursor logic can be replaced with one line of T-SQL

SELECT @studentsEnrolled = COUNT(*) FROM courseEnrollment

Which one do you think will perform faster?

What is more evil than a cursor?

If cursors are evil, then what is more evil than a cursor? Nested cursors of course, especially three nested cursors. Here is an example of some horrible code where a cursor is not needed

CREATE PROCEDURE [dbo].[sp_SomeMadeUpName]


SET @SomeDate =  CONVERT(CHAR(10),getDate(),112)

EXEC sp_createSomeLinkedServer @SomeDate,@SomeDate,12


FROM SomeTable
WHERE getDate() BETWEEN SomeStart and SomeEnd

OPEN SomeCursor



 SET @sql = ''
 SET @sql = @sql + N'DECLARE @Date DATETIME, @Value FLOAT' + char(13) + char(13)
 SET @sql = @sql + N'DECLARE curData CURSOR FOR' + char(13)
 SET @sql = @sql + N'SELECT * ' + char(13)
 SET @sql = @sql + N'FROM OPENQUERY(LinkedServerName,''SELECT Date,' + RTRIM(@SomeID) + ' FROM SomeTable'')' + char(13) + char(13)
 SET @sql = @sql + N'OPEN curData' + char(13) + char(13)
 SET @sql = @sql + N'FETCH NEXT FROM curData INTO @Date,@Value' + char(13)
 SET @sql = @sql + N'WHILE @@FETCH_STATUS = 0' + char(13)
 SET @sql = @sql + N'BEGIN' + char(13)
 SET @sql = @sql + N'INSERT INTO SomeTAble' + char(13)
 SET @sql = @sql + N'VALUES(''' + @SomeID + ''',@Date,@Value)' + char(13)
 SET @sql = @sql + N'FETCH NEXT FROM curData INTO @Date,@Value' + char(13)
 SET @sql = @sql + N'END' + char(13)
 SET @sql = @sql + N'CLOSE curData' + char(13) + char(13)
 SET @sql = @sql + N'DEALLOCATE curData' 

 PRINT @sql + char(13) + char(13)

 EXEC sp_ExecuteSQL @sql



CLOSE SomeCursor

Why the need of looping over the list of IDs? Join with the linked server and do all this in 3 lines of code

I have seen some really horrible code with nested cursors, one example I saw was when someone needed to sum up some data, he created 3 nested cursor..the first one to loop over years, the second to loop over months and the third to loop over days.  This thing ran forever

All you need is to do a simple group by... for example

;WITH cte AS (SELECT DATEADD(dd,number,'20190101') AS TheDate,number
FROM master..spt_values WHERE type = 'p')

SELECT SUM(number) as Value,
YEAR(TheDAte)as TheYear, 
MONTH(TheDate) AS TheMonth 
FROM cte
ORDER BY TheYear,TheMonth


Replacing one evil with another

If you are using while loops instead of cursors then you really have not replaced the cursor at all, you are still not doing a set based operation, everything is still going row by row. Aaron Bertrand has a good post here, no need for me to repeat the same Bad Habits to Kick : Thinking a WHILE loop isn't a CURSOR

Loops in triggers

Usually you will find a loop in a trigger that will call some sort of stored procedures that needs to perform some kind of task for each row affected by the trigger. In the Triggers, when to use them, when not to use them post I already explained how to handle this scenario

Loops in stored procedures

Here is where you will find the biggest offenders. All you have to do to find the procs with cursors is run the following piece of code

SELECT * FROM sys.procedures 

For while loops, just change the '%DECLARE%%cursor%' part to '%while%'
Look at those procs and investigate if you can rewrite them using a SET based operation instead

Loops for maintenance purposes

One of the few times I will use cursors or while loops is if I need to get information about tables or databases where I have to get information from a stored procedure.
Here is an example

CREATE TABLE #tempSpSpaceUsed (TableName VARCHAR(100),
        Rows INT,
        Reserved VARCHAR(100),
        Data VARCHAR(100),
        IndexSize VARCHAR(100),
        Unused VARCHAR(100))
INTO #temp
FROM sys.tables 

DECLARE @loopid INT = 1, @maxID int = (SELECT MAX(id) FROM #temp)
WHILE @loopid <= @maxID

SELECT @TableName = name FROM #temp WHERE id = @loopid
INSERT #tempSpSpaceUsed
 EXEC('EXEC sp_spaceused ''' + @TableName + '''')
SET @loopid +=1

SELECT * FROM #tempSpSpaceUsed
Of course this can be simplified as well, you can just run this instead of the while loop and run the output

EXEC sp_spaceused ''' + [name] + ''''
FROM sys.tables

Finally if you need more cursor goodness, take a look at these three posts which are also mentioned in the wiki article

Tuesday, January 01, 2019

SQL Server Maintenance

It has been a while since I wrote some of my best practices posts. I decided to revisit these posts again to see if anything has changed, I also wanted to see if I could add some additional info.

In this post  we are going to look at SQL Server maintenance

Just like with a car or a house, you need to do maintenance on databases as well. SQL Server has gotten better over the years, there are less knobs you need to turn out of the box but maintenance is still required.

In this post I will be looking at some stuff that you need to be aware of. Some of the things I will mention can be thought of as maintenance as well as regular checks. Think of a DBA as a car mechanic, instead of an oil change, tune up or checking the tire pressure, the DBA will check index fragmentation, run DBCC CHECKDB and make sure you have enough space for the database to grow for the next predetermined period.

The things I will cover in this post are: fragmentation of indexes, free drives space, space in filegroups, running DBCC CHECKDB and finally making sure that you have the latest source code of your objects in a source control system.

Check fragmentation of indexes

A lot of time your index will get fragmented over time if you do a lot of updates or inserts and deletes.

Now instead of rolling your own solution, you should take a look at some of them that are out there and used by many people. Take a look at SQL Server Index and Statistics Maintenance by Ola Hallengren, You can also get the scripts from Github here: https://github.com/olahallengren/sql-server-maintenance-solution

Check that your database is healthy by running DBCC CHECKDB

What does DBCC CHECKDB do? Here is the explanation from Books On Line
Checks the logical and physical integrity of all the objects in the specified database by performing the following operations:

  • Runs DBCC CHECKALLOC on the database.
  • Runs DBCC CHECKTABLE on every table and view in the database.
  • Runs DBCC CHECKCATALOG on the database.
  • Validates the contents of every indexed view in the database.
  • Validates link-level consistency between table metadata and file system directories and files when storing varbinary(max) data in the file system using FILESTREAM.
  • Validates the Service Broker data in the database.
So how frequent should you be running DBCC CHECKDB? Ideally you should be running DBCC CHECKDB as frequent as possible, do you want to find out that there is corruption when it is very difficult to fix since two weeks have passed or do you want to find out the same day so that you can fix the table immediately.

Paul Randal who worked on DBCC CHECKDB has a whole bunch of blog posts about DBCC CHECKDB, the posts can be found here http://www.sqlskills.com/blogs/paul/category/checkdb-from-every-angle.aspx

Make sure that you have enough space left on the drives

Running out of space on a drive is not fun stuff, suddenly you can't insert any more data into your tables because no new pages can be allocated. If you have tools in your shop like cacti then this is probably already monitored. If you don't have any tools then either get a tool or roll your own. Here is how you can get the free space fo the drives with T-SQL

CREATE TABLE #FixedDrives(Drive CHAR(1),MBFree INT)

INSERT #FixedDrives
EXEC xp_fixeddrives

SELECT * FROM #FixedDrives

Here is the output for one of my servers

Drive MBFree
C 6916  -- System
D 28921 -- Apps
L 52403 -- Log
M 4962  -- System databases
T 86208 -- Temps
U 71075 -- User databases 
V 212075-- User databases 

Here is a simple way of using T-SQL to create a SQL Agent job that runs every 10 minutes and will send an email if you go below the threshold that you specified. This code is very simple and is just to show you that you can do this in T-SQL. You can make it more dynamic/configurable by not hard-coding the drives or thresholds

CREATE TABLE #FixedDrives(Drive CHAR(1),MBFree INT)

INSERT #FixedDrives
EXEC xp_fixeddrives

FROM #FixedDrives

FROM #FixedDrives

DROP TABLE #FixedDrives

IF @MBFreeD < 30000 OR @MBFreeE < 10000
      DECLARE @Recipients VARCHAR(8000)
   SELECT @Recipients ='SomeGroup@SomeEmail.com'
  DECLARE @p_body AS NVARCHAR(MAX), @p_subject AS NVARCHAR(MAX), @p_profile_name AS NVARCHAR(MAX)

  SET @p_subject = @@SERVERNAME + N'  Drive Space is running low'
  SET @p_body = ' Drive Space is running low <br><br><br>' + CHAR(13) + CHAR(10) + 'Drive D has ' 
  + CONVERT(VARCHAR(20),@MBFreeD) + ' MB left <br>' + CHAR(13) + CHAR(10) + 'Drive E has ' 
  + CONVERT(VARCHAR(20),@MBFreeE) + ' MB left'

  EXEC msdb.dbo.sp_send_dbmail
     @recipients = @Recipients,
     @body = @p_body,
     @body_format = 'HTML',
     @subject = @p_subject

Make sure that you have enough space left for the filegroups

In the Sizing database files I talked about the importance of sizing database files. Just like you can run out of hard drive space, you can also fill up a file used by SQL Server. here is query that will tell you how big the file is, how much space is use and how much free space is left. You can use a query like this to alert you before you run out of space

  CONVERT(DECIMAL(12,2),ROUND((a.size-FILEPROPERTY(a.name,'SpaceUsed'))/128.000,2)) ,
 NAME = LEFT(a.NAME,35),
 dbo.sysfiles a

Have the latest scripts of all your objects

You might say that you have all the code for your objects in the database. What if you want to go back to the version of the proc from 3 days ago, is it really easier to restore a 800 GB backup from 3 days ago just to get the stored proc code?

Of course not, make sure that you have DDL scripts of every object in source control, your life will be much easier.
I only touched on a couple of points here, some of the things mentioned here will also show up in the proactive notifications post in a couple of days. There is much more to maintenance than this, keep informed and make sure you understand what needs to be done.

Some more repos for you to use

Take a look at the GitHub repositories mentioned in this post: Five great SQL Server GitHub repos that every SQL Server person should check out  There are some good ones like dbatools and tigertoolbox