CS/IT Tutorials Collections

Best Practices For Sql Server Database Files And Tempdb Considerations

Pinterest LinkedIn Tumblr
The SQL database is a critical component that many people rely on heavily. SQL Server is a relational database whose primary purpose is to store and retrieve data for other programs.

Databases used to be flat, which meant they stored data in long log text files called tab-delimited files. Those very files encapsulated information from multiple sources about objects, resources, and employees in each entry. Although the data was organized as a record, it was inconsistent. Teams found it difficult and time-consuming to search for specific information in the file and create customized reports because the data was organized in a sparse format.

Some of the best practices for SQL server and tempdb considerations will be discussed today to gain a better understanding. Explore for SQL Server Tutorial for more information.

Best practices for SQL server database files:

1. Security:

You’ll need a trustworthy person, someone who can grant SQL Server management permissions. This is critical because you don’t want everyone conducting surveillance around and gaining access to corporate information. I would also recommend encrypting your backups and restricting access to only a few users. You do not want that to happen to the data you’ve backed up. Spyware is very common these days, so make sure you have a backup in a hidden and secure facility that only you can access. Maintain it on a SharePoint site rather than a mounted drive, and restrict access to your DBAs.

If anybody keeps having DBA permissions, kindly notify them that they’re in charge of system administration and data backup. See how they react to that concept. I’m sure no developer would agree to such a deal.

I recommend using Windows Authentication rather than just your system admin account while connecting to SQL Servers (SA). Likewise, avoid using the SA account to connect to SQL Server web applications.

Make a habit of changing passwords and auditing file logins regularly to spot any irregularities. Changing passwords too frequently is a lot of work—you should do it now and then, but don’t go overboard and modify it once every few hours.

 2. Management of data and logs:

Assure you have a separate drive for MDF and LDF log files when installing the SQL server. Instead of storing them on different partitions, each one requires its own physical drive. That is the most crucial piece of advice I can offer you.

They tend to run slower as they’re on a relatively similar physical disc. That’s because the disc writes will happen simultaneously; in fact, this same procedure will switch from writing log files to writing data files at random intervals.

It’s not a big deal if you’re using an SSD drive; many folks still use physical discs that run on plates. This is due to the fact that SQL databases can be quite large, and SSD drives can be quite costly when you need tens of terabytes.

If the databases aren’t very active, people must be on the identical drive. In reality, this is all based on the quantity and activity of databases.

3. Diminishing Database file:

The standard best practice is to prevent dwindling database files; however, there are times when shrinking is beneficial and times when it is simply not recommended. It wouldn’t recommend depleting data files on active databases, as once the shrinking starts, all other database transactions are halted. Databases that aren’t critical, including the non-production database, development, or UAT, can be shrunk. If you ever need to free up disc space, you could even shrink them. For instance, shrinking it is perfectly acceptable if you’re working with a small database and only need a simple transaction log rather than a full one.

If you feel compelled to scale down the files to save space, resist. Maintain your composure. It’s a terrible idea. Instead, make a database expansion strategy. Downsizing can extensively separate the index and overturn the sequence of the clustered index, causing the database to perform poorly. If you must, shrink files, but not the database!

4. Default auto-growth settings should be changed:

The transaction would then start taking as often as they grow + the transaction if the truncation you operate over through the database necessitates more space than you will have free, so the truncation may time out if the auto grows too large. So it is recommended That the MB set increments rather than percentages. After all course, there is really no silver bullet when it comes to what figures you can access because the size of the databases is so important.

Supervise the expansion of the databases over time and prepare ahead when setting the auto to grow MB value. This will allow you to track the size of your database over time and predict database size based on historical snapshots.

5. Datafiles are split across multiple drives:

Make absolutely sure you have a current backup of the Virtual machine and database throughout in case the documented procedure fails. Another very popular DB configuration with such a single data file and only one log file designated to a database.

For example, if the initial database was 10GB in size with 2GB of free space, and you fully intend to split it across four drives, every one of the new databases should not be any larger than 2GB at first. This process will help us reduce the size of the DB’s original file while also ensuring that the data is evenly distributed across new data files.

6. Documentation for SQL Server:

Creating thorough SQL Server documentation is difficult, tedious, and time-consuming. However, SQL Server documentation is required. Now, allow us to explain why this is so critical. If nothing else, having an internal IT infrastructure is always beneficial.

Proper documentation consists of :

  • information on all of your SQL servers
  • details about diverse database files, as well as a list of databases and their sizes
  • SQL configuration options in general
  • server supervisors, user privileges, and database owners’ information
  • results are compared to your settings to best practices in the industry.

7. Backup:

Before you begin, you must have a structured plan, at the very least with the most important databases, even in the worst-case scenario. Automatic backups can always be scheduled, but significant company information should be managed and backed up by manual processes as well.

Data loss is unthinkable, so you should back up your data regularly. Take extra precautions when backing up transaction logs. If you don’t, the transaction log might grow in a full mode based on database activity, allowing you to reestablish to any point in time using only the transaction log. The tool noted will also display the last date of the database backup.

8. Patch SQL Servers regularly:

Regularly installing service packs, security fixes, and total combined updates should be your top priority. So that you can do that, users must stay current on the latest patches and the SQL Server support cycle. After that, you can go over how to patch a SQL Server. The tool I mentioned will assist you in checking server patches and will inform you if your server is out of date and what CUs or SPs you need to install.

Tempdb considerations:

Because database file location affects I/O performance so much, you should consider functional changes to tempdb when planning your main data placement strategy. Because tempdb is the most dynamic database here on the system and must be the fastest, its performance has a significant impact on overall system performance.

Tempdb, along with all other databases, is made up of primary data and log files. User objects and internal objects are stored in tempdb. It even has two different versions. A version store is a set of data pages that hold the data rows needed to support row versioning features. The following are the two version stores:

  • Data manipulation transactions in tempdb, which use glimpse or read dedicated create row versions. separation levels for row versioning
  • Data modification transactions create row versions in tempdb for features like online index operations, Multiple Active Result Sets (MARS), and AFTER triggers.

The mentioned huge feature set, which generates user and internal objects or edition stores, has been added to tempdb:

  • Query
  • Triggers
  • Read committed screenshots and snapshot isolation
  • There are several active result sets (MARS)
  • Creating an online index
  • Table variables, table-valued functions, and temporary tables
  • Check DBCC
  • Parameters for Large Objects (LOB)
  • Cursor
  • Event notification and service broker
  • Variables XML and Large Object (LOB)
  • Notifications for queries
  • Email from a database
  • Creating an index
  • Definable functions

It is possible to achieve good performance by putting the tempdb database on a solely devoted and incredibly fast I/O subsystem. To enhance scalability, a lot of progress has been made on tempdb internals.

To ensure that tempdb is properly sized and capable of handling the needs of your enterprise system, you should be doing at least some capacity planning.

At the very least, do the following:

1. Take into account the size of your current tempdb.

2. Keep an eye on tempdb while running the processes that impact it most. The following query returns the five most frequently executed tasks that use tempdb:

most frequently executed tasks that use tempdb

While keeping an eye on tempdb, rebuild the index of your largest table online. Because this process now takes place in tempdb, don’t be amazed if such a number rolls out to be two times the table size.

The following query should be run at regular intervals to keep track of tempdb size. It is suggested that this be done at least once a week. This query determines and expresses the amount of tempdb space used by internal objects, free space, version store, and user objects (in kilobytes):

query that determines and expresses the amount of tempdb space

This query’s output could look something like this:

query's output

Taking into account the preceding results, the following steps must be taken when configuring tempdb:

  • Pre-allocate space for tempdb files relying on your test results, but keep autogrow facilitated in case tempdb runs out of room, and SQL Server crashes.
  • As a general rule, create one tempdb data file per CPU or processor core per SQL Server instance, with a maximum of eight data files.
  • Ascertain that tempdb is set to the simple recovery model, which allows for space recovery.
  • Set auto growth to set size in the range of 10% of tempdb’s original size.
  • Place tempdb on a high-performance, dedicated I/O system.
  • Use immediate database file creation if the SQL Server (MSSQLSERVER) Service account does not have admin privileges.


The practices listed above should assist you in achieving your goals. SQL server data has proven to be valuable for most software developers when developing software applications. Nonetheless, the world has changed dramatically. SQL is a fantastic database management system. A SQL’s future potential is limitless; it can be found in retail, finance, healthcare, science and technology, government, and, in short, anywhere. A database is required for all businesses to store their data. Waiting about 10-15 years will make things easier. Backup systems, corrupt practices, repair, and restoration assistance will become less of an issue.


Sai Priya Ravuri is a Digital Marketer, and a passionate writer, who is working with MindMajix, a top global online training provider. She also holds in-depth knowledge of IT and demanding technologies such as Business Intelligence, Machine Learning, Salesforce, Cybersecurity, Software Testing, QA, Data analytics, Project Management and ERP tools, etc.

Comments are closed.