Change Tracking May Leads to Blocking and Dead Locks

Changing Tracking: May not be a viable solution on OLTP databases.

CT (change tracking) primarily used for Synchronizing the data.

As per Books Online: It tracks what rows are changed. The latest data can be obtained directly from the table that is being tracked. It’s a light weight method to track the changes, unlike triggers which needs programming with additional tables and cleanup process."

But it’s not light weight in terms of the problems raised with blocking.

Change tracking itself creates few internal tables without having any control over them to maintain or troubleshoot during the issues raised with performance.

In this scenario, we wants to synchronize the tables.

Steps involved in synchronize 2 tables using change tracking are

1.     Populate the Target table with initial full set of data from the source table

2.     Identify the changes on the source tables using change tracking

3.     Synchronize both the tables with the changes (insert,update,delete) happens on the source table to the destination table

There are many methods available viz..Triggers, Time Stamp Tracking, Log shipping, Replication, Mirroring (with snapshot) etc. and each method has its own limitations

But if the solution has using change tracking then beware of the problem with deadlocks.

"A deadlock occurs when two or more tasks permanently block each other by each task having a lock on a resource which the other tasks are trying to lock.

Image: Dead Lock Trace Captured from Profiler

Steps:

1.     Create a Test Database

2.     Create a test table with test data

3.     Enable Change Tracking on database

4.     Enable Change Tracking on Table

5.     Run Update Statements

6.     Catch Deadlocks

7.     Conclusion

8.     Alternate solutions

1.     Create a Test Database

Create database TestDB Go Use TestDB Go

2.     Create a test table with test data

USE TestDB GO create table [Customer](     IID INT IDENTITY (1,1) NOT NULL PRIMARY KEY CLUSTERED, [CustomerID] [int] NOT NULL , [PersonID] [int] NULL, [StoreID] [int] NULL, [TerritoryID] [int] NULL, [AccountNumber]  varchar(30), [rowguid] [uniqueidentifier] ROWGUIDCOL  NOT NULL, [ModifiedDate] [datetime] NOT NULL) GO insert into [Customer] SELECT * FROM AdventureWorks2008.Sales.Customer

3.     Enable Change Tracking on database

ALTER DATABASE TESTDB SET CHANGE_TRACKING = ON (CHANGE_RETENTION = 7 DAYS, AUTO_CLEANUP = ON);

4.     Enable Change Tracking on Table

USE TestDB GO ALTER TABLE customer  enable CHANGE_TRACKING WITH (TRACK_COLUMNS_UPDATED = Off);

5.      

6.     Run Update Statements

-- open a new session(1) and run the below statement while 1=1 BEGIN insert into [Customer] SELECT * FROM AdventureWorks2008.Sales.Customer END; -- open a 2nd new session(2) and run the below statement while 1=1 BEGIN     update Customer set PersonID =customerid END;

7.     Catch Deadlocks

select XEventData.XEvent.value('(data/value)[1]', 'varchar(max)') as DeadlockGraph FROM (select CAST(target_data as xml) as TargetData from sys.dm_xe_session_targets st join sys.dm_xe_sessions s on s.address = st.event_session_address where name = 'system_health' and target_name = 'ring_buffer') AS Data CROSS APPLY TargetData.nodes ('//RingBufferTarget/event') AS XEventData (XEvent) where XEventData.XEvent.value('@name', 'varchar(4000)') = 'xml_deadlock_report'

8.     Conclusion:

make sure before implementing change tracking on the tables with high volume of concurrent insert and updates. Change tracking creates few internal tables which are the main sources of causing these deadlocks.
During every modification on source table , it creates one record in the internal change tracking table. So , If Transaction A inserting records every 1 second and at the same time Transaction B is updating the records than the concurrent blocking will be placed on the Internal tables between these 2 transactions which leads to deadlocks.

Choose this option wisely or write your own custom code using other features where you can take full control.
The major challenge involved in synchronizing the tables are identifying the modifications (updates) than the inserts/deletes as inserts/deletes can be identified by using an anchor table by comparing the primary keys in any data warehouse environment.

Major challenge here is identify the modifications, we have other options mentioned below to identify the modified rows where each has its own limitations. Choose your options wisely

i)       Binary checksum

ii)      Timestamp

iii)     Non-key attributes with in the table to keep track of the changes by the application like dt_modified

iv)    Triggers

v)     CDC – change data capture

vi)    CT – change tracking

vii)   T-Replication

viii)  3^rd party tools

Also refer the article written by Brentozar at the below link. His article pretty much covered about change tracking.

http://www.brentozar.com/archive/2014/06/performance-tuning-sql-server-change-tracking/

backups on shared database instances with out maintenance plans

 

Managing database backups on shared database instances without maintenance plan wizard.

Download Code

Problem

As the number of databases with different recovery models at different schedules exists in shared database instances, so do the challenges related managing the databases backups. Brent clearly described in his post about the few of the challenges http://www.brentozar.com/archive/2012/12/backups-gone-bad-the-danger-of-differentials/

Maintenance plans are very helpful to create and schedule the backups, but it has very limited features where you can’t control the backups of all the databases in shared database instances. 

	Maintenance Plans	This Solution
To Include the databases to the existing Plan	manual	dynamic query as a parameter "select name from sys.databases where name in ('db_1','db_2')
Copying/Moving the backups to secondary location		supports. set the job parameter "I_Optional_RemoteBkpPath]" with the secondary location path
Track the status (backup,copies,deleted) of each backup including its secondary locations		supports. select * from BKPS_DB_AUTO_LOG
delete the old backups based on dependencies instead age	based on age	based on your own query.
restart the job from the point of failure		supports
dashboard monitoring		supports. create your own dashboard using the bkps_db_auto_log table.

Below framework/solution addresses all the problems listed above. You can modify the same solution as per your requirement to control multiple instances with multiple databases. and you can configure and check whether the backup jobs are configured as per your DRP requirements.

1. Backups all the databases based on the recovery model. by default it takes full database backups of all the databases and takes t-log backups only if the recovery model is full or bulk-logged.
2. no manual intervention required to include the databases to take full backups
3. track and moves the backup files from local server to secondary locations . once all the backups are completed in a batch.
4. deletes the old backup files in sequence (full + dependent differential and t-log backups) instead based on the no.of days old at the secondary location.
5. a daily report to ensure that all the backups are in place

Solution

Description :

Steps are very self descriptive.

1. Get DB List : gets the list of the databases from the output of the input variable I_SqlDBList  to include in local variable "dblist".
2. Get Next Execution ID : insert a record in dbo.BKPS_DB_AUTO_LOG with an auto increment number during each execution
3. Loop All DB List : loops through all the databases in the list
1. creates a directory for each database in the local folder
2. backup the database
4. get files to move from local to remote location : If the optional variable @I_Optional_RemoteBkpPath is not null , then it assigns the list of the files to move to a local variable " moveFileList"
5. Move and delete old files : loops through the list of the backup files assigned in step 1
1. moves one file at a time in a loop
2. get old file list to delete : get the list of the OLD files which are depends on the file which is just moved
3. Loop old backups : loops through one file at a time to delete
1. delete old backups : delete and log the status of the file deletion in the log table

Input variables Description
set the variables where name starts with I_ , rest are local variables

Variable Name	Description
I_BkpExtension	Backup extension. Don’t enter “.”
I_BkpIsCompressed	Enter 1 to compress the backups
I_BkpIsVerify	Enter 1 to verify the backup
I_BkpType	Enter the backup type Full /Diff/Log
I_LocalBkpPath	Enter local database backup path. If you want to take the backups directly to a remote location then enter the remote location name
I_Optional_RemoteBkpPath	This is optional. Enter remote location to move the local backups to this location.
I_Optional_SqlRetention	This is Optional. Enter the select query to include the list of the backup files to delete. Eg: below query produce the list of old backups except recent,full +recent differential + all log backups after recent diff backup. select files_to_delete from vw_keep_recent_full_nd_diff
I_SqlDBList	Enter the select query to list the databases to take the backup. Eg: below value includes the backups X,Y,Z select name as dblist from master.sys.databases where name in ('X','Y','Z') You can filter the database based on your own criteria. For e.g.: To take the log backups of all databases which are in full recovery model , set the variable value with “select name as dblist from sys.databases where recovery_model_desc='full' “

SSIS Flow Steps Description

  Database Objects

1. BKPS_DB_AUTO_LOG (Table) : it’s a tracking table , to store the status of the backup executions and location of the backup files
2. fn_split_files  : user defined function
3. BKP_INS_BKP_STATS(procedure) : to insert the status of the backup operation in tracking table
4. BKP_DBS_MULTI (procedure) : to take the backups
5. vw_keep_recent_full_nd_diff(view) : this view contains the all the old files to delete from the remote location. You can change this view as per your retention plan.

Steps to configure the job on your lab server

1. Download all the files

2. Save the BKPS_DB_AUTO.dtsx as an external file in c drive

3. execute all the scripts (1-6)  in MSDB database
Note: you should review and modify the view vw_keep_recent_full_nd_diff to delete the old backups , I have provided this view for simulation purpose only .

4. Create and Configure SQL Server agent Job
after executing the script "6.create Job BKP_DBS_AUTO" , it create a job with name "BKP_DBS_AUTO"

f you already has a job with the same name then modify the first line in the script with a new name for the job
Go to SQL server agent , right click on “"BKP_DBS_AUTO" ” job , open the first step and modify the location of the SSIS package

1.

2 Click on [Set value] tab on same screen and modify the values for all variables as below example. Refer the section “variable description” in this article for the description of these variables

2. Execute the job, after successful completion of the job , you will see the records in msdb.dbo.BKPS_DB_AUTO_LOG table

3. Execute the job several times , and check the status column in the same table. it deletes all the old backup files and keep only the recent one. If you don’t want to delete the old backups then don’t pass any value to the variable I_Optional_SqlRetention

Next:

To create multiple jobs , you can script out the existing job and change the variable values

More on deleting old backups:

Attached download files has a view “vw_keep_recent_full_nd_diff” , this view refers the tracking table “bkp_file_locations” and produce the list of the remote old backup files to delete except the recent full ,recent differentials and all log files after the recent differential backup.

I have created a view to produce the list of the old backup files to delete to reuse it.

create view vw_keep_recent_full_nd_diff as --select BFL.DBNAME,BFL.IID ,RECENT.IID ,BFL.BKP_TYPE ,RECENT.BKP_TYPE select BFL.BKP_REMOTE_PATH from BKP_file_locations BFL LEFT outer join (select MAX(IID)AS IID,DBNAME,BKP_TYPE from bkp_file_locations where bkp_type <> 'LOG' GROUP BY DBNAME,BKP_TYPE ) recent on recent.dbname = BFL.dbname AND recent.BKP_TYPE = BFL.BKP_TYPE AND recent.IID = BFL.IID WHERE RECENT.IID IS NULL AND BFL.status = 'moved'

To test what database file are deleted. Run the below query, check the difference in status columns.

select iid,DBNAME,bkp_type,status from BKP_file_locations order by dbname

Below is the output for the databases X and Y

sql server partition table online from non partition table

step by step process to partition an existing table to a partitioned table



steps :

1. create table orders
2. populate test data with past 12 months to current date
3. lock the table till the switch happens , create partition function
4. create partition scheme
5. create partition table
6. create check constraints , check constraints are needed to switch the table to new partition
7. switch the existing table to temp table and rename the temp to newly
   partitioned table and finally drop the existing table
8. insert new test data
9. check the rows on partition

------------------------------------------------------- -- Create test database ------------------------------------------------------- create database poc GO use POC ------------------------------------------------------- -- 1 Create the new Orders table. ------------------------------------------------------- CREATE TABLE Orders (iid int identity(1,1),   nage int,   CreatedDate datetime ) GO ------------------------------------------------------- -- 2 populate test data with past 12 months to current date ------------------------------------------------------- -- insert the orders from 2012 december to till this month insert into Orders select 10,'2012-dec-01' insert into Orders select 10,'2013-jan-01' insert into Orders select 10,'2013-jan-15' insert into Orders select 10,'2013-jan-31' insert into Orders select 10,'2013-feb-01' insert into Orders select 10,'2013-feb-15' insert into Orders select 10,'2013-feb-28' insert into Orders select 10,'2013-mar-01' insert into Orders select 10,'2013-mar-15' insert into Orders select 10,'2013-mar-31' insert into Orders select 10,'2013-apr-01' insert into Orders select 10,'2013-apr-15' ------------------------------------------------------- -- 3 create partition function -- create new table with partition. -- create partition starting next month, since you want to archive all the old data after 1 month or so ------------------------------------------------------- BEGIN TRAN select * from orders with (holdlock , tablockx) where 1=0 create partition function PF_MONTHLY_ORDERS (datetime) as RANGE RIGHT for values( '2001-01-01', -- from this date , get the minimum date -1 month '2013-05-01', -- to this date , get maxdate + 1 month to < this date it goes to 1st partition. if the current month is April(4th) '2013-06-01', -- add one month for next 12 months '2013-07-01', '2013-08-01','2013-09-01', '2013-10-01', '2013-11-01', '2013-12-01' ) ------------------------------------------------------- -- 4 create partition scheme ------------------------------------------------------- create PARTITION SCHEME PSCH_MONTHLY_ORDERS  as partition PF_MONTHLY_ORDERS  all to ([primary]) go ------------------------------------------------------- -- 5 create partition table -- ansi setting must be on as mentioned below ------------------------------------------------------- SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO SET ANSI_PADDING OFF GO CREATE TABLE [Orders_temp] (iid int identity(1,1),   nage int,   CreatedDate datetime ) ON PSCH_MONTHLY_ORDERS(CreatedDate) GO SET ANSI_PADDING OFF GO ------------------------------------------------------- -- 6 create constraints on both the tables -- check constraints are needed to switch the tables , with out this its not possible to move the data ------------------------------------------------------- alter table orders add constraint chkPrtn1 check (CreatedDate >='2001-01-01'  and CreatedDate is not null) alter table orders add constraint chkPrtn2 check (CreatedDate < '2013-05-01' and CreatedDate is not null) alter table orders_temp add constraint chkPrtn3 check (CreatedDate >='2001-01-01'  and CreatedDate is not null) alter table orders_temp add   constraint chkPrtn4 check (CreatedDate < '2020-01-01' and CreatedDate is not null) -------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------- -- 7  switch the data to newly partioned table ------------------------------------------------------- alter table orders switch partition 1 to orders_temp partition 2 alter table orders_temp drop constraint chkPrtn3 alter table orders_temp drop constraint chkPrtn4 exec sp_rename orders ,orders_x exec sp_rename orders_temp ,orders select * from Orders COMMIT TRAN -- !! ROLLBACK OR COMMIT THE TRANSACTION -- !! CHECK AND DROP THE TABLE CpAuditLog_x -------------------------------------------------------------------------------------------------------------------------- -- 9 insert new test data insert into Orders select 10,'2013-apr-30' -- this will goes to 2 partition insert into Orders select 10,'2013-may -30' -- this will goes to 3 partition -------------------------------------------------------------------------------------------------------------------------- --  10 check the rows on partitition SELECT [CreatedDate],         $partition.PF_MONTHLY_ORDERS([CreatedDate])             AS 'Partition Number' FROM Orders SELECT $partition.PF_MONTHLY_ORDERS([CreatedDate])             AS 'Partition Number'     , min([CreatedDate]) AS 'Min Order Date'     , max([CreatedDate]) AS 'Max Order Date'     , count(*) AS 'Rows In Partition' FROM Orders GROUP BY $partition.PF_MONTHLY_ORDERS([CreatedDate]) SELECT sprv.value AS [Value], sprv.boundary_id AS [ID] ,* FROM sys.partition_functions AS spf INNER JOIN sys.partition_range_values sprv ON sprv.function_id=spf.function_id WHERE (spf.name=N'PF_MONTHLY_ORDERS') ORDER BY [ID] ASC ------ full schema query DECLARE @TableName sysname = 'orders'; SELECT p.partition_number, fg.name, p.rows,* FROM sys.partitions p     INNER JOIN sys.allocation_units au     ON au.container_id = p.hobt_id     INNER JOIN sys.filegroups fg     ON fg.data_space_id = au.data_space_id WHERE p.object_id = OBJECT_ID(@TableName) --------------other meta data queries DECLARE @TableName sysname = 'orders'; SELECT distinct p.partition_number, fg.name, p.rows,sprv.value AS [Value] FROM sys.partitions p     INNER JOIN sys.allocation_units au     ON au.container_id = p.hobt_id     INNER JOIN sys.filegroups fg           ON fg.data_space_id = au.data_space_id     INNER JOIN sys.destination_data_spaces dds           ON fg.data_space_id = dds.data_space_id     INNER JOIN sys.partition_schemes ps ON ps.data_space_id = dds.partition_scheme_id     INNER JOIN sys.partition_functions AS spf  ON spf.function_id = ps.function_id     INNER JOIN sys.partition_parameters pp ON pp.function_id=spf.function_id     INNER JOIN sys.partition_range_values sprv ON sprv.function_id=pp.function_id and sprv.parameter_id= pp.parameter_id WHERE p.object_id = OBJECT_ID(@TableName) and (spf.name=N'PF_MONTHLY_ORDERS') SELECT sprv.value AS [Value], sprv.boundary_id AS [ID] ,* FROM sys.partition_functions AS spf INNER JOIN sys.partition_range_values sprv ON sprv.function_id=spf.function_id WHERE (spf.name=N'PF_MONTHLY_ORDERS') ORDER BY [ID] ASC

SQL Server I/O Errors List

If you are frequently encountering the I/O errors like below  , and to analyze when exactly these are happening you need
to export all the io errors into a table , so that you can sort based on date and time to see the trends.
below script will export all the io errors from error log in sql server into a temp table #io_errors
output :

use tempdb create table #IO_ERRORS(LogDate datetime,ProcessInfo varchar(30),text varchar(max)) declare @error_msg  varchar(2000) = 'requests taking longer than 15 seconds to complete on file' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 0, 1,@error_msg  , '', NULL, NULL, N'desc' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 1, 1, @error_msg  , '', NULL, NULL, N'desc' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 2, 1, @error_msg  , '', NULL, NULL, N'desc' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 3, 1, @error_msg  , '', NULL, NULL, N'desc' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 4, 1, @error_msg  , '', NULL, NULL, N'desc' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 5, 1, @error_msg  , '', NULL, NULL, N'desc' insert into #IO_ERRORS EXEC master.dbo.xp_readerrorlog 6, 1, @error_msg  , '', NULL, NULL, N'desc' select LogDate,physical_name, OCCURANCES= cast(SUBSTRING(TEXT,CHARINDEX('SQL Server has encountered ',TEXT)+27, CHARINDEX('occurrence(s) of I/O requests taking longer than ',TEXT)-CHARINDEX('SQL Server has encountered ',TEXT)-27) as int) --,CAST(logdate as DATE) as dt, DATEPART(HH,logdate   ) as h ,DATEPART(MINUTE,logdate   ) as minu from #IO_ERRORS a JOIN sys.master_files b on a.text like  '%'+b.physical_name +'%' order by 3 desc

sql server update outdated statistics

 

Though the auto update statistics are ON, SQL Server wont update the statistics unless it reaches certain threshold  http://support.microsoft.com/kb/195565

If you are using SQL Server 2008 R2 or later version then the below link will be helpful. http://support.microsoft.com/kb/2754171?wa=wsignin1.0

But If you are using the 2005 or 2008 (not R2)  or don't want to enable the trace in R2 as mentioned in above link then below procedure will helps :

1- Update the statistics when the table got modified with in 24 hours delay from last statistics update
2-Updates only the tables which are modified instead all the tables in the database.

CREATE PROC PR_UPDATE_STATS AS with cte as ( select SCHEMA_NAME(b.schema_id) as tbl_schema,OBJECT_NAME(b.object_id  ) as tbl_name, STATS_DATE(s.object_id, s.stats_id) as stat_dt,last_user_update, (select MAX(last_user_dt) from (values (last_user_seek),(last_user_scan),(last_user_lookup)) as all_val(last_user_dt)) as last_seeks_lkps_scans_datetime from sys.dm_db_index_usage_stats  c  join sys.tables b on b.object_id =  c.object_id join sys.stats s on b.object_id = s.object_id ) select 'UPDATE STATISTICS '+ tbl_name  +';' as cmd, tbl_schema , tbl_name , MAX(last_user_update) as last_udpate , MAX(stat_dt) as last_stat_dt ,max(last_seeks_lkps_scans_datetime) as last_select  into #t from cte group by tbl_schema , tbl_name having datediff(dd,MAX(stat_dt),MAX(last_user_update)   ) >=1 -- stats which are outdated for 1 day   declare @cmd varchar(256) declare cur cursor for select cmd from #t open cur     fetch next from cur into @cmd while @@fetch_status =  0     begin     exec (@cmd)         fetch next from cur into @cmd     continue     end                close cur deallocate cur

 

This is a simple and straight forward solution to update the statistics based on capturing what tables are updated since the last update statistics time. But if you want to control update statistics based on threshold (no.of updates ) then you have to collect the metrics with how many updates are happening on each table through collecting sys.dm_db_index_usage_stats on daily basis and modify the stored procedure by adding a parameter where last_user_update > @threshold_limit.

sql server when was the database table accessed last time

-- this works starting 2008 onwards with cte as ( select (select min(last_user_dt) from (values (last_user_update),(last_user_seek),(last_user_scan),(last_user_lookup)) as all_val(last_user_dt)) as dt ,DB_NAME(database_id) as dbName from sys.dm_db_index_usage_stats ) select min(dt) as since from cte where dbName = 'NORTHWIND' -- to check when was the database accessed last time. use the below. source : https://www.mssqltips.com/sqlservertip/3171/identify-sql-server-databases-that-are-no-longer-in-use/

sql server script to know when the table accessed last time

If the database has lot many unwanted tables which are not being used by any of the applications and if you want to identify the ununused objects , so that you can remove or move those objects to a separate database , then the system view sys.dm_db_index_usage_stats (sql server 2008) contains the information with when the table accessed last time.
the script SQL-1 will produce the list of the tables which are being used and non-used as well. the records in the column [recent_datetime] contains the list of tables which are non been used since the server started.
Note : the counters on this DMV starts / refresh when the server is restarted.

-- Query-1 with cte_recent as ( select SCHEMA_NAME(B.schema_id) +'.'+object_name(b.object_id) as tbl_name, (select MAX(last_user_dt) from (values (last_user_seek),(last_user_scan),(last_user_lookup)) as all_val(last_user_dt)) as access_datetime FROM sys.dm_db_index_usage_stats a right outer join sys.tables b on a.object_id =  b.object_id ) select tbl_name,max(access_datetime) as recent_datetime  from cte_recent group by tbl_name order by recent_datetime desc , 1

output

To keep a track of the tables which are non been used , its better to store the results in a separate table and merge the table before restart the sql server or as per your own schedule as the SQL Server installed on cluster then you probably schedule it as a job to wait for an year before making the decision to remove or move those tables to a different database.
create a table TABLES_LAST_ACCESSED with 2 columns as below script

--Query-2 create table TABLES_LAST_ACCESSED (TBL_NAME VARCHAR(256),recent_datetime DATETIME)

sync the table with the script (Query-3) below

-- Query-3 with cte_recent as ( select SCHEMA_NAME(B.schema_id) +'.'+object_name(b.object_id) as tbl_name, (select MAX(last_user_dt) from (values (last_user_seek),(last_user_scan),(last_user_lookup)) as all_val(last_user_dt)) as access_datetime FROM sys.dm_db_index_usage_stats a right outer join sys.tables b on a.object_id =  b.object_id ) MERGE TABLES_LAST_ACCESSED AS target USING (select tbl_name,max(access_datetime) as recent_datetime  from cte_recent group by tbl_name) AS source ON (Target.tbl_name = Source.tbl_name) when matched then         UPDATE SET Target.recent_datetime = Source.recent_datetime when not matched then         INSERT (tbl_name,recent_datetime)VALUES (source.tbl_name,source.recent_datetime);

query any table by using select , insert , update or delete and execute the below statement. the tables which you accessed recently should be in the top.

SELECT * FROM  TABLES_LAST_ACCESSED ORDER BY 2 DESC

For more information on this DMV http://msdn.microsoft.com/en-us/library/ms188755.aspx




SQL Server 2005 : http://stackoverflow.com/questions/711394/how-do-you-find-the-last-time-a-database-was-accessed