Backup frequency

A backup strategy is NOT a technical decision, it is the result of applying technical expertise to business decisions regarding RPO (Recovery Point Objective) and RTO (Recovery Time Objective).

 

For example, if you are backing up a 5T database, it doesn’t really matter if you are backing up frequently if you can’t restore within your own defined SLA (Service Level Agreement). On top of that, you must (and usually will) need to decided what amount of data is a reasonable loss (if any). That is to say, if you are backing up your database daily, and taking incremental backups hourly, can you afford to lose the 30-minutes worth of data that will be lost, on average, should the servers crash.

 

In addition, you need to discuss the types of disaster you are backing up for. For example, if you don’t have an air-gapped backup, you may be paying ransomware at some point in your future.

 

For a very small (few hundred gig) system, you might  backup:

 

Full backups every 24 hours,

Log Backups every 15  minutes,

 

And realize that you have decided that a 7.5-minute loss is something that you live with if everything hits the fan.

 

If you have a multi-terabyte database, you may find that it takes hours to backup the “standard” way and that you need to use other tools, hardware, or perhaps add differential backups to the mix.

 

And in either case, make very sure that you have:

 

1. Made sure that the backup is free of corruption
2. Instructed the system  group to sweep the backups off the target disks
3. Tested your restore process from the remote storage.

 

Backups are a 4-hour lecture; or a 30-minute white board session with executives. This is only a start.

What queries are running at the moment?

select
text
FROM
sys.dm_exec_requests er
INNER JOIN sys.sysprocesses sp ON er.session_id = sp.spid
CROSS APPLY sys.dm_exec_sql_text(er.sql_handle)as qt
WHERE
session_Id > 50 and -- exclude system stuff
text like '%create index%' -- find something we care about
and session_id <> @@spid -- exclude this session

When is an index redundant?

Recently I was called into a shop that had a 4 terabyte database, who claimed they didn’t have anywhere near that much data. Upon investigation, they were correct. They had about 100 gig of data, and about 3.9 T of indexes.

Is this too much? Well, in order to make that kind of judgement call there are several factors to consider:

1. Are the indexes interfering with writes (i.e. is the number of indexes having a negative factor on insert / update performance)?
2. Are the indexes actually being used?
3. Are any indexes redundant?

The first one (write performance) can be identified by checking length of the write queues; almost any monitoring tool can get you this information.

The second can be identified with a script you can find on the web which counts the number of times an index has been accessed since the last time the server was booted. Of course, you want to the server to have been up long enough that the system tables are aware of weekend / month end processing. If you can’t find one readily, you can use the one I have on my blog:

Mssqlperformance.blogspot.com

The final one is the tool I used to reduce the database size for that customer from 4T to 300M without even doing “what’s being used” analysis… just by eliminating redundant indexes.

Many people find a script on the web (I have one in the blog mentioned above) which will look at the system tables and recommend missing indexes. The problem is, it will give you 6 similar indexes rather than one combined index. You have to do this yourself. If you automatically add all the indexes the server suggests… well, you can end up with 4T of indexes for 100G of data.

So how do we know what’s redundant?

Hopefully, you are old enough to remember a phone book. A phone book is an index on last name, first name, middle name. But what other indexes might that be used for? A lookup on just last name, first name could use the same phone book (and often does), as would a lookup just to find last name.

So if you have three indexes:

1. Last name, first name, middle name
2. Last name, first name
3. Last name

You can visually look at these and KNOW that the last 2 are redundant because the first index can do the work that either of the second two could do.

As you start looking at data management view (DMV) information, you often get index recommendations along these lines:

1. Create index idx on TableA (Column1)
2. Create index idx on TableA (Column1) include (ColumnA, ColumnB)
3. Create index idx on TableA (Column1, Column2) include (ColumnC, ColumnD)

Sometimes you’ll get 8 or 10 of these.

So, index 1 can easily be eliminated because index 2 covers it.

Index 2 can be combined into index 3 like this:

1. Create index idx on TableA (Column1, Column2) include (ColumnA, ColumnB , ColumnC, ColumnD)

So… in summary, you can pull index recommendations from the DMVs, and they are good recommendations, but review them with an eye for detail before automatically applying them.

List indexes with keys by table

Note: Code borrowed from:

https://stackoverflow.com/questions/765867/list-of-all-index-index-columns-in-sql-server-db

... and modified to suit my needs.

SELECT
  schema_name(schema_id) as SchemaName, OBJECT_NAME(si.object_id) as TableName, si.name as IndexName,
  (CASE is_primary_key WHEN 1 THEN 'PK' ELSE '' END) as PK,
  (CASE is_unique WHEN 1 THEN '1' ELSE '0' END)+' '+
  (CASE si.type WHEN 1 THEN 'C' WHEN 3 THEN 'X' ELSE 'B' END)+' '+  -- B=basic, C=Clustered, X=XML
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,1,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,2,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,3,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,4,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,5,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,6,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  '' as 'Type',
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,1) as Key1,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,2) as Key2,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,3) as Key3,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,4) as Key4,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,5) as Key5,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,6) as Key6
FROM sys.indexes as si
LEFT JOIN sys.objects as so on so.object_id=si.object_id
WHERE index_id>0 -- omit the default heap
  and OBJECTPROPERTY(si.object_id,'IsMsShipped')=0 -- omit system tables
  and not (schema_name(schema_id)='dbo' and OBJECT_NAME(si.object_id)='sysdiagrams') -- omit sysdiagrams
ORDER BY SchemaName ,TableName , Key1, Key2, Key3, Key4, Key5, Key6

This version excludes tables with only one index:

SELECT

  schema_name(schema_id) as SchemaName, OBJECT_NAME(si.object_id) as TableName, si.name as IndexName,

  (CASE is_primary_key WHEN 1 THEN 'PK' ELSE '' END) as PK,

  (CASE is_unique WHEN 1 THEN '1' ELSE '0' END)+' '+

  (CASE si.type WHEN 1 THEN 'C' WHEN 3 THEN 'X' ELSE 'B' END)+' '+  -- B=basic, C=Clustered, X=XML

  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,1,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+

  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,2,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+

  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,3,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+

  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,4,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+

  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,5,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+

  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,6,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+

  '' as 'Type',

  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,1) as Key1,

  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,2) as Key2,

  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,3) as Key3,

  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,4) as Key4,

  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,5) as Key5,

  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,6) as Key6

FROM sys.indexes as si

LEFT JOIN sys.objects as so on so.object_id=si.object_id

WHERE index_id>0 -- omit the default heap

  and OBJECTPROPERTY(si.object_id,'IsMsShipped')=0 -- omit system tables

  and not (schema_name(schema_id)='dbo' and OBJECT_NAME(si.object_id)='sysdiagrams') -- omit sysdiagrams

  and so.object_id in (select object_id  from sys.indexes group by object_id having count(object_id) > 1)

ORDER BY SchemaName ,TableName , Key1, Key2, Key3, Key4, Key5, Key6

Excessive compile times

This code was brought to my attention by an associate, Valentin Iamandi, who mentioned it was from Jonathan Kehayias ... we have a client who may be victim of a SQL Server compilation bug (3 seconds compile time, 75ms execution time, this is to identify other potential victims.

-- Find high compile resource plans in the plan cache

SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;

WITH XMLNAMESPACES 

(DEFAULT 'http://schemas.microsoft.com/sqlserver/2004/07/showplan')

SELECT TOP 10

CompileTime_ms,

CompileCPU_ms,

CompileMemory_KB,

qs.execution_count,

qs.total_elapsed_time/1000 AS duration_ms,

qs.total_worker_time/1000 as cputime_ms,

(qs.total_elapsed_time/qs.execution_count)/1000 AS avg_duration_ms,

(qs.total_worker_time/qs.execution_count)/1000 AS avg_cputime_ms,

qs.max_elapsed_time/1000 AS max_duration_ms,

qs.max_worker_time/1000 AS max_cputime_ms,

SUBSTRING(st.text, (qs.statement_start_offset / 2) + 1,

(CASE qs.statement_end_offset

WHEN -1 THEN DATALENGTH(st.text)

ELSE qs.statement_end_offset

END - qs.statement_start_offset) / 2 + 1) AS StmtText,

query_hash,

query_plan_hash

FROM

(

SELECT 

c.value('xs:hexBinary(substring((@QueryHash)[1],3))', 'varbinary(max)') AS QueryHash,

c.value('xs:hexBinary(substring((@QueryPlanHash)[1],3))', 'varbinary(max)') AS QueryPlanHash,

c.value('(QueryPlan/@CompileTime)[1]', 'int') AS CompileTime_ms,

c.value('(QueryPlan/@CompileCPU)[1]', 'int') AS CompileCPU_ms,

c.value('(QueryPlan/@CompileMemory)[1]', 'int') AS CompileMemory_KB,

qp.query_plan

FROM sys.dm_exec_cached_plans AS cp

CROSS APPLY sys.dm_exec_query_plan(cp.plan_handle) AS qp

CROSS APPLY qp.query_plan.nodes('ShowPlanXML/BatchSequence/Batch/Statements/StmtSimple') AS n(c)

) AS tab

JOIN sys.dm_exec_query_stats AS qs

ON tab.QueryHash = qs.query_hash

CROSS APPLY sys.dm_exec_sql_text(qs.sql_handle) AS st

ORDER BY CompileTime_ms DESC

OPTION(RECOMPILE, MAXDOP 1);

Unused indexes -- updated script

DECLARE @cmd VARCHAR(6000)

DECLARE @TempIndexDrop TABLE (
       [Database_Name] [nvarchar](128) NULL,
       [Object_Name] [nvarchar](128) NULL,
       [Index_Name] [sysname] NULL,
       [Drop_Statement] [nvarchar](283) NULL,
       [user_seeks] [bigint]  NULL,
       [user_scans] [bigint]  NULL,
       [user_lookups] [bigint]  NULL,
       [SpaceSavedMeg] decimal (12,2)  null
) 

SET @cmd = 'USE [?] ;
select db_name(database_id) as ''Database_Name'', object_name(i.object_id) as ''Object_Name'',
i.name as ''Index_Name''
, ''USE ['' + db_name(database_id) + '']; drop index [''  + i.name + ''] on ['' + sch.name + ''].['' + obj.name + '']'' as ''Drop_Statement'',
user_seeks, user_scans, user_lookups,
(select sum(s.used_page_count)*8./1024. from sys.dm_db_partition_stats s where i.object_id = s.object_id and i.index_id = s.index_id)
from sys.dm_db_index_usage_stats ius join sys.indexes i
on ius.index_id = i.index_id and ius.object_id = i.object_id join sys.objects obj
on ius.object_id = obj.object_id join            sys.schemas sch
on            obj.schema_id = sch.schema_id
where      
database_id = db_id(''?'') and
user_seeks + user_scans + user_lookups = 0
and i.name not like ''pk%''
and ''?'' not in (''master'', ''model'',''tempdb'',''msdb'')

order by       object_name(i.object_id),       i.name'

INSERT INTO @TempIndexDrop 
(
[Database_Name]
,[Object_Name]
,[Index_Name] 
,[Drop_Statement] 
,[user_seeks]
,[user_scans]
,[user_lookups]
,[SpaceSavedMeg]
)
Exec sp_msforeachdb @cmd

SELECT SUM(SpaceSavedMeg) as 'Potential Storage Savings' FROM  @TempIndexDrop

SELECT *
FROM @TempIndexDrop
ORDER BY [Database_Name], [Object_Name]

Missing indexes -- revised script

Thanks to Darren Myher (https://DarrenMyher.com) for a few intelligent tweaks as well as some useful formatting for this revised version.

--Use this SQL Script to identify indexes that if added could help improve performance.

--IMPORTANT: Do not BLINDLY implement the script recommendations.  For example, if the SAME TABLE is mentioned multiple times, then

--           it may be better to COMBINE the recommendations of multiple statements into a single one.

--           see later posting on this blog for detail

DECLARE @Edition varchar(50)

SET @Edition = CONVERT(varchar(50), SERVERPROPERTY('Edition'))

DECLARE @bEnterpriseEdition bit

IF (LEFT(@Edition, 9) = 'Enterpris' OR LEFT(@Edition, 9) = 'Developer')

SET @bEnterpriseEdition = 1

SELECT

      'CREATE INDEX IX_'

            + REPLACE(obj.Name, ' ', '') +'_'

            + replace(replace(replace (equality_columns, '[', ''),']',''),', ','_')

            + ' ON '

            + sch.name COLLATE SQL_Latin1_General_CP1_CI_AS + '.' + QuoteName(obj.name)

            + ' ('

            + equality_columns

            + CASE WHEN inequality_columns IS NULL

                   THEN ''

                   ELSE ',' +  inequality_columns

              END

            + ')'

            + CASE WHEN included_columns IS NOT NULL

                   THEN ' INCLUDE (' + ISNULL(included_columns,'') + ') '

                   ELSE ''

              END

            + CASE WHEN @bEnterpriseEdition = 1

                   THEN ' WITH (online = ON)'

                   ELSE ''

              END

            + ' -- ' + CONVERT (varchar, avg_user_impact) + '% anticipated impact'

      FROM sys.dm_db_missing_index_details mid

        JOIN sys.dm_db_missing_index_groups mig

            ON mid.index_handle = mig.index_handle

        JOIN sys.dm_db_missing_index_group_stats migs

            ON migs.group_handle = mig.index_group_handle

        JOIN sys.objects obj

            ON obj.object_id = mid.object_id

        JOIN sys.schemas sch

            ON obj.schema_id = sch.schema_id

      WHERE

        database_id = db_id()

        AND avg_user_impact > 60

        AND equality_columns IS NOT NULL

      ORDER BY

        obj.name

        , equality_columns

        , inequality_columns

        , included_columns