Click here to Skip to main content
Click here to Skip to main content

SQL SERVER – Database Worst Practices – New Town and New Job and New Disasters

, 22 Jul 2011
Rate this:
Please Sign up or sign in to vote.
This blog post is written in response to the T-SQL Tuesday hosted by Amit Banerjee. I know Amit personally and very big fan of his community activities. I read his blog, read his comments, follow his tweets, and the most importantly reach out to him when I feel like talking SQL. Amit has selected a

This blog post is written in response to the T-SQL Tuesday hosted by Amit Banerjee. I know Amit personally and very big fan of his community activities. I read his blog, read his comments, follow his tweets, and the most importantly reach out to him when I feel like talking SQL.

Amit has selected a very interesting subject – Best Practices. When I read the subject, I can directly relate this subject to my real world interactions. I have seen so many developers and DBAs making fatal mistakes in the early days of their career. Most of the time, they get another chance because they are new to the job or that particular task, which they are doing for the first time. However, such excuses cannot be made for experts and senior professionals. They are expected perform and carry out the right practices always.

Instead of writing best practices, I am going to write about few of the worst practices which I have quite often seen in the real world. Some of them are really bad that I often wonder how those who still use them were able to continue on the job so far. Let us go over few of the worst practices I have observed in the industry.

My log file is growing too big – I truncated the log without taking backup

This has to be the numero uno of worst practices. Every time I come across such a statement, I almost miss a heartbeat. Before I continue, let me confess that at the beginning of the career, I have practiced the same. After a while, I had learned it the hard way. This is never a good practice; truncating the log file is not an option. The reason why I rate this as the worst practice: this one mistake can make database impossible to recover. When someone truncates the log file without taking backup, there is no chance of recovery.

Here are a couple of articles which I have written on how to prevent log file from growing too big.

I shrink my database daily to regain the space

This is one of the popular worst practices. I have seen administrators shrinking the database at the end of the day to gain the space only to lose it the very next day. Shrinking is a VERY BAD operation. It increases fragmentation, reduces the performance, and wastes the resources. I strongly advise not to do it.

Here are few articles I had earlier written on this subject.

Clustered Index makes the table to sort every time. I do not have clustered index on any table

For an OLTP system, Index is very important and clustered index is the most important index (in my opinion). Clustered index forces order to the table and removes the ‘forwarding records’ problem from the database. Personally, I consider tables without clustered index performance to be unacceptable?. In my OLTP system, I always recommend that all tables should have clustered index.

Here is a quick script that can help in identifying a table without clustered index in the database.

TempDB is not important; so I will keep it on my slow drive

Personally, I have tremendous respect for TempDB. Even though it resets every time when the server restarts, this is a single most important database that is shared among all the other databases in the system. This database is used for sorting, temporary objects, triggers, row version and in other operations. Keeping it on the slow drive is not the solution, but reality is that it will just create many performance-related problems in the overall system. If your TempDB is becoming full, move it to the another drive.

Here are a few blog posts I have written on TempDB.

I am confused between Full, Differential, and Log Backup

Inability to understand the proper recovery model is another worst practice. I have people restoring many differential backups while restoring the database. I quite often see that log file backup interval is so huge it is more than differential backup interval. There are so many factors which can lead to disaster and data loss, leading to people to look for a new job in a new town at times. If you are confused regarding what is tail log backup, then stop, and learn from online books before implementing the backup strategy. Even if you are not responsible for implementing the backup strategy, I would still suggest you to read how to carry out proper backup as you never know when it will land heavily on your job card!

Here are few interesting write-ups on this subject on this blog.

There are many more practices I can write, but I think these five are the top worst practices. Feel free to post your opinions and suggestions.

Reference: Pinal Dave (http://blog.SQLAuthority.com)

License

This article, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

About the Author

pinaldave
Founder http://blog.SQLAuthority.com
India India
Pinal Dave is a Microsoft Technology Evangelist (Database and BI). He has written over 2200 articles on the subject on his blog at http://blog.sqlauthority.com. Along with 8+ years of hands on experience he holds a Masters of Science degree and a number of certifications, including MCTS, MCDBA and MCAD (.NET). He is co-author of two SQL Server books - SQL Server Programming, SQL Wait Stats and SQL Server Interview Questions and Answers. Prior to joining Microsoft he was awarded Microsoft MVP award for three continuous years for his contribution in community.
Follow on   Twitter

Comments and Discussions

 
-- There are no messages in this forum --
| Advertise | Privacy | Mobile
Web04 | 2.8.140721.1 | Last Updated 22 Jul 2011
Article Copyright 2011 by pinaldave
Everything else Copyright © CodeProject, 1999-2014
Terms of Service
Layout: fixed | fluid