How does MySQL store data - mysql

I looked around Google but didn't find any good answers. Does it store the data in one big file? What methods does it use to make data access quicker than just reading and writing to a regular file?

This question is a bit old but I decided to answer it anyway since I have been doing some digging on the same. My answer is based on the linux file system. Basically mySQL stores data in files in your hard disk. It stores the files in a specific directory that has the system variable "datadir". Opening a mysql console and running the following command will tell you exactly where the folder is located.
mysql> SHOW VARIABLES LIKE 'datadir';
+---------------+-----------------+
| Variable_name | Value |
+---------------+-----------------+
| datadir | /var/lib/mysql/ |
+---------------+-----------------+
1 row in set (0.01 sec)
As you can see from the above command, my "datadir" was located in /var/lib/mysql/. The location of the "datadir" may vary in different systems. The directory contains folders and some configuration files. Each folder represents a mysql database and contains files with data for that specific database. below is a screenshot of the "datadir" directory in my system.
Each folder in the directory represents a MySQL database. Each database folder contains files that represent the tables in that database. There are two files for each table, one with a .frm extension and the other with a .idb extension. See screenshot below.
The .frm table file stores the table's format. Details: MySQL .frm File Format
The .ibd file stores the table's data. Details: InnoDB File-Per-Table Tablespaces
That’s it folks! I hope I helped someone.

Does it store the data in one big file?
Some DBMSes store the whole database in a single file, some split tables, indexes and other object kinds to separate files, some split files not by object kind but by some storage/size criteria, some can even entirely bypass the file system, etc etc...
I don't know which one of these strategies MySQL uses (it probably depends on whether you use MyISAM vs. InnoDB etc.), but fortunately, it doesn't matter: from the client perspective, this is a DBMS implementation detail the client should rarely worry about.
What methods does it use to make data access quicker them just reading and writing to a regular file?
First of all, DBMses are not just about performance:
They are even more about safety of your data - they have to ensure there is no data corruption even in the face of a power cut or a network failure.1
DBMSes are also about concurrency - they have to arbiter between multiple clients accessing and potentially modifying the same data.2
As for your specific question of performance, relational data is very "susceptible" to indexing and clustering, which is richly exploited by DBMSes to achieve performance. On top of that, the set-based nature of SQL lets the DBMS choose the optimal way to retrieve the data (in theory at least, some DBMSes are better at that than the others). For more about DBMS performance, I warmly recommend: Use The Index, Luke!
Also, you probably noticed that most DBMSes are rather old products. Like decades old, which is really eons in our industry's terms. One consequence of that is that people had plenty of time to optimize the heck out of the DBMS code base.
You could, in theory, achieve all these things through files, but I suspect you'd end-up with something that looks awfully close to a DBMS (even if you had the time and resources to actually do it). So, why reinvent the wheel (unless you didn't want the wheel in the first place ;) )?
1 Usually though some kind of "journaling" or "transaction log" mechanism. Furthermore, to minimize the probability of "logical" corruption (due to application bugs) and promote code reuse, most DBMSes support declarative constraints (domain, key and referential), triggers and stored procedures.
2 By isolating transactions and even by allowing clients to explicitly lock specific portions of the database.

Technically everything is a "file" including folders.. your entire hard drive is giant file. Having said that, yes relational databases, MySQL included store data in a Data file on the hard drive. The difference between a Database and writing/reading to a file is apples and oranges. Databases provide a structured way to store and search/retrieve data in a way you could never replicate by just reading and writing to a file.. Unless you wrote your own db of course..
hope that helps.

When you store data in a flat file, it is compact and efficient to read sequentially, but there is no fast way to access it randomly. This is especially true of variable-length data such as documents, names or strings. To allow for fast random access, most databases store information in a single file using a data structure called a B-Tree. This structure allows for insert, deletion, and search to be fast, but it can use up to 50% more space than the original file. Typically, however, this is not an issue as disk space is cheap and larger, while the primary tasks usually require fast access.
For more information:
http://en.wikipedia.org/wiki/B-tree
Looking carefully into the MySQL docs, we find that indices may be optionally set to "BTREE" or "HASH" type. Inside a single MySQL file, multiple indices are stored which may use either data structure.
Although safety and concurrency are important, these are not WHY databases exist, but added features. The very first databases exist because it is not possible to randomly access a sequential file containing variable length data.

Related

Unable to understand how ibd file encryption is helping secure the data in MySql

I am using MySql 8.0.11 and using keyring_file plugin, I have encrypted a particular table, say t1, in my database.
When I check the contents of t1.ibd file I can see that the contents are encrypted after successful encryption. But, I continue to see the table contents using query (select * from t1) even if the contents in the ibd file are encrypted.
So, does it mean that the encryption works on ibd files only (which contains the data and the indexes) but I will continue to see the table contents without any issue if I have the database credentials?
UPDATE
I read a couple of comments and would like to add the below questions to clarify my original query:
After encrypting the ibd files, if a hacker hacks into my system on which the database is hosted, the hacker would be able to see the actual data. So, how has encrypting the ibd files helped me secure the data?
An .ibd file contains all the data and all the indexes for a table. ("Tablespaces" can contain multiple tables; the principles are the same.)
With the plugin (etc), SELECT automagically goes through the decryption, making the encryption 'transparent'. Nevertheless it is real. You did have to do something to start the program, correct? That was when the 'key' was loaded into RAM for use.
Encrypting .ibd files protects you from one threat: someone grabs (or copies) your disk drive.
But beware. There are temp tables, binlogs, other logs, etc, that may or may not be encrypted. They temporarily hold some of the data. Early versions of MySQL encryption failed to include some of these.
The AES functions let you en/decrypt individual strings (such as one column in one row at a time). But this leaves it up to you to protect the en/decryption key. Or at least never have it on disk as plaintext.
Read about "encryption at rest" versus "encryption in flight". Encrypting the files is "at rest". A smart hacker will attack your code so that he can run a SELECT after the credentials have been loaded.
With or without encryption, "SQL injection" (qv) is a well established way to hack into the data, and even into the filesystem. The protection for this comes, for example, comes by validating/escaping/etc data that comes from an HTML <form>. Encrypting files is no protection against this.
I have given you a short list of threats against your data. The real list is much longer. You need to find such a list and decide which ones you are willing to invest in protecting against. Security is not simple.

What is difference between storing data in a blob, vs. storing a pointer to a file?

I have a question about the blob data type in MySQL.
I read that the data type can be used to store files. I also read that an alternative is to store the file on disk and include a pointer to its location in the database (via a varchar column).
But I'm a little confused because I've read that blob fields are not stored in-row and require a separate look-up to retrieve its contents. So is that any different than storing a pointer to a file on the file system?
I read that the data type can be used to store files.
According to MySQL manual page on Blob, A BLOB is a binary large object that can hold a variable amount of data.
Since it's a data type specific to store binary data it's common to use it to store files in binary format, being storing image files a very common use on web applications.
For web applications this would mean that you would first need to convert your file into binary format and then store it, and every time you need to retrieve your file you would need to do the reverse process of converting them back to it's original format.
Besides that, storing large amount of data in your db MAY slow it down. Specially in systems that are not dedicated only to host a database.
I also read that an alternative is to store the file on disk and include a pointer to its location in the database
Bearing in mind all above considerations a common practice for web applications is to store your files elsewhere than your MySQL and then simply store it's path on your database. This approach MAY speed up your database when dealing with large amount of data.
But I'm a little confused because I've read that blob fields are not stored in-row and require a separate look-up to retrieve its contents.
In fact that would depend on what storage engine you are using since every engine treats data and stores it in different ways. For the InnoDB engine, which is suited for relational database you may want to read this article from MySQL Performance blog on how the blob is stored in MySQL.
But in abstract, on MySQL 5 and forward the blob is stored as following:
Innodb stores either whole blob on the row page or only 20 bytes BLOB pointer giving preference to smaller columns to be stored on the page, which is reasonable as you can store more of them.
So you are probably thinking now that the right way to go is to store them as separate file, but there are some advantages of using blob to store data, the first one (in my opinion) is the backup. I manage a small server and I had to create another subroutine only to copy my files stored as paths to another storage disk (We couldn't afford to buy a decent tape backup system). If I had designed my application to use blobs a simple mysqldump would be everything that I needed to backup my whole database.
The advantage of storing blobs for backups are better discussed on this post where the person who answered had a similar problem than mine.
Another advantage is security and the easiness of managing permission and access. All the data inside your MySQL server is password protected and you can easily manage permissions for your users about who access what and who doesn't.
In a application which relies on MySQL privileges system for authentication and use. It's certain a plus since it would be a little harder for let's say an invader to retrieve an image (or a binary file like a zipped one) from your disk or an user without access privileges to access it.
So I'd say that
If you gonna manage your MySQL and all the data you have in it and must do regular backups or intend to change or even consider a future change of OS, and have a decent hardware and optimized your MySQL to it, go for BLOB.
If you will not manage your MySQL (as in a web host for example) and doesn't intend to change OS or make backups, stick with varchar columns pointing to your files.
I hope it helped. Cheers
If you store data is BLOB field, you are making it part of your object abstraction.
BLOB advantages:
Should you want to remove row with BLOB, or remove it as part of master/slave table relationship or maybe the whole table hierarchy, your BLOB is handled automatically and has same lifetime as any other object in database.
Your scripts do not have a need to access anything but database to get everything they require. In many situations, having direct file access open whole can of worms on how to bypass access or security restrictions. For example, with file access, they may have to mount filesystems which contain actual files. But with BLOB in database, you only have to be able to connect to database, no matter where you are.
If you store it in file and file is replaced, removed or no longer accessible, your database would never know - in effect, you cannot guarantee integrity. Also, it is difficult to reliably support multiple versions when using files. If you use and depend on transactions, it becomes almost impossible.
File advantages:
Some databases handle BLOBs rather poorly. For example, while official BLOB limit in MySQL is 4GB, but in reality it is only 1MB in default configuration. You can increase this to 16-32MB by tweaking both client and server configuration to increase MySQL command buffer, but this has a lot of other implications in terms of performance and security.
Even if database does not have some weird size limits, it always will have some overhead in storing BLOB compared to just a file. Also, if BLOB is large, some databases do not provide interface to access blob piece by piece, or stream it, which can be large impediment for your workflow.
In the end, it is up to you. I typically try to keep it in BLOB, unless this creates unreasonable performance problems.
Yes, MySQL blobs that don't fit within the same page as a row get stored on overflow pages Note that some blobs are small enough that they're stored with the rest of the row, like any other column. The blob pages are not adjacent to the page their row is stored on, so they may result in extra I/O to read them.
On the other hand, just like with any other page type, blob pages can occupy memory in the InnoDB buffer pool, so reading the blobs subsequently is very fast even if they are on separate pages. Files can be cached by the operating system, but typically they're read from disk.
Here are a few other factors that may affect your decision:
Blobs are stored logically with a row. This means if you DELETE the row, the associated blob is deleted automatically. But if you store the blob outside the database, you end up with orphaned blob files after you delete rows from the database. You have to do manual steps to find and delete these files.
Blobs stored in the row also follow transaction semantics. For instance, a new blob or an updated blob is invisible to other transactions until you commit. You can also roll back a change. Storing blobs in files outside the database makes this a lot harder.
When you back up a database containing blobs, the database is a lot bigger of course, but when you backup, you get all the data and associated blobs in one step. If you store blobs externally, you have to back up the database and also back up the filesystem where you store blob files. If you need to ensure that the data and blobs are captured from one instant in time, you pretty much need to use some kind of filesystem snapshots.
If you use replication, the only automatic way of ensuring the blobs get copied to the replication slave automatically is to store blobs in the database.
Filesystem access will be faster than through the database. Blobs columns have some disadvantages in terms of indexing/sorting etc, which you could do with your filename column if you wished to in the future.
The database can also grow quickly with large blobs and then tasks like backing up become slower. I would go with a file location in database with the physical storage on the file system.
The better approach is to store your file in the filesystem folder and point to their paths through a varchar field in the database. One of the drawbacks of saving files in the database is slowing it or reducing its performance.

How do databases physically store data on a filesystem?

I need to know how data from databases is stored on a filesystem. I am sure, that different databases use different ways of storing data, but I want to know what the general rule is (if there is one), and what can be changed in settings of a particular DB.
How is the whole database stored? In one big file or one file per table?
What if a table is enormous? Would it be split into few files?
What is typical size of file in that case?
The answer to this question is both database dependent and implementation dependent. Here are some examples of how data can be stored:
As a single file per database. (This is the default for SQL Server.)
Using a separate file system manager, which could be the operating system. (MySQL has several options, with names like InnoDB.)
Using separate files for each table. (If we consider Access a database.)
As multiple physical files, spread across multiple file systems, but represented as a single "file". (HIVE, for instance, that uses a parallel file system to store the data.)
However, these are the default configurations. Real databases typically let you split the data among multiple physical devices. SQL Server and MySQL call this partitions. Oracle calls this table spaces. These are typically set up by knowledgeable DBAs who understand the performance requirements of the system.
The final questions are easy to answer, though. Most databases give you the option of either growing the databases as space is needed or giving the database a fixed (or fixed maximum) size. I have not encountered a database engine that will split the underlying data into multiple files automatically, although it is possible that newer column oriented databases (such as Vertica) do something similar.

What database should I use to store timestamped pictures (or generally BLOBS)?

I need a database that can store a large number of BLOBS. The BLOBs would be picture files and would also have a timestamp and a few basic fields (size, metrics, ids of objects in other databases, things like that), but the main purpose of the database is to store the pictures.
We would like to be able to store the data in the database for a while, in the order of few months. With the data coming in maybe every few minutes, the number of BLOBs stored can grow quite quickly.
For now (development phase) we will be using a MySQL for this. I was wondering if MySQL is a good direction to go, in terms of:
Being able to store binary data efficiently
Scalability
Maintenance requirements.
Thanks,
MySQL is a good database, and can handle large data sets. However, there is a great benefit in making your whole database fit into RAM, in such case all database-related activity will be much faster. By putting large and seldom-accessed objects into your database, you're making this harder.
So, I think a combined approach is the best:
Save only metadata in the database, and save the files on disk as-is. Better to hash the directories if you're talking about 100,000 of files, then save file under the name of an index field in your database. E.g. such directory structure:
00/00001.jpg
00/00002.jpg
00/00003.jpg
....
....
10/10234.jpg
10/10235.jpg
In this case, your directories won't have too many files, and accessing the files is fast and easy. Of course if your database server is distributed/redundant, things get more interesting, any such approach may or may not be warranted, depending on the load, redundancy/fail over requirements, etc.
I suggest to store images on hard disk and in your mysql implementation maintain the metadata of your image including the filename (maybe). So your script can easily pick it up from your local hard drive.
For Reading & storing files, hard disk and most modern OS are really good at it. So I believe mysql is not going to solve anything here.

How does MySQL store rows on disk? [duplicate]

I looked around Google but didn't find any good answers. Does it store the data in one big file? What methods does it use to make data access quicker than just reading and writing to a regular file?
This question is a bit old but I decided to answer it anyway since I have been doing some digging on the same. My answer is based on the linux file system. Basically mySQL stores data in files in your hard disk. It stores the files in a specific directory that has the system variable "datadir". Opening a mysql console and running the following command will tell you exactly where the folder is located.
mysql> SHOW VARIABLES LIKE 'datadir';
+---------------+-----------------+
| Variable_name | Value |
+---------------+-----------------+
| datadir | /var/lib/mysql/ |
+---------------+-----------------+
1 row in set (0.01 sec)
As you can see from the above command, my "datadir" was located in /var/lib/mysql/. The location of the "datadir" may vary in different systems. The directory contains folders and some configuration files. Each folder represents a mysql database and contains files with data for that specific database. below is a screenshot of the "datadir" directory in my system.
Each folder in the directory represents a MySQL database. Each database folder contains files that represent the tables in that database. There are two files for each table, one with a .frm extension and the other with a .idb extension. See screenshot below.
The .frm table file stores the table's format. Details: MySQL .frm File Format
The .ibd file stores the table's data. Details: InnoDB File-Per-Table Tablespaces
That’s it folks! I hope I helped someone.
Does it store the data in one big file?
Some DBMSes store the whole database in a single file, some split tables, indexes and other object kinds to separate files, some split files not by object kind but by some storage/size criteria, some can even entirely bypass the file system, etc etc...
I don't know which one of these strategies MySQL uses (it probably depends on whether you use MyISAM vs. InnoDB etc.), but fortunately, it doesn't matter: from the client perspective, this is a DBMS implementation detail the client should rarely worry about.
What methods does it use to make data access quicker them just reading and writing to a regular file?
First of all, DBMses are not just about performance:
They are even more about safety of your data - they have to ensure there is no data corruption even in the face of a power cut or a network failure.1
DBMSes are also about concurrency - they have to arbiter between multiple clients accessing and potentially modifying the same data.2
As for your specific question of performance, relational data is very "susceptible" to indexing and clustering, which is richly exploited by DBMSes to achieve performance. On top of that, the set-based nature of SQL lets the DBMS choose the optimal way to retrieve the data (in theory at least, some DBMSes are better at that than the others). For more about DBMS performance, I warmly recommend: Use The Index, Luke!
Also, you probably noticed that most DBMSes are rather old products. Like decades old, which is really eons in our industry's terms. One consequence of that is that people had plenty of time to optimize the heck out of the DBMS code base.
You could, in theory, achieve all these things through files, but I suspect you'd end-up with something that looks awfully close to a DBMS (even if you had the time and resources to actually do it). So, why reinvent the wheel (unless you didn't want the wheel in the first place ;) )?
1 Usually though some kind of "journaling" or "transaction log" mechanism. Furthermore, to minimize the probability of "logical" corruption (due to application bugs) and promote code reuse, most DBMSes support declarative constraints (domain, key and referential), triggers and stored procedures.
2 By isolating transactions and even by allowing clients to explicitly lock specific portions of the database.
Technically everything is a "file" including folders.. your entire hard drive is giant file. Having said that, yes relational databases, MySQL included store data in a Data file on the hard drive. The difference between a Database and writing/reading to a file is apples and oranges. Databases provide a structured way to store and search/retrieve data in a way you could never replicate by just reading and writing to a file.. Unless you wrote your own db of course..
hope that helps.
When you store data in a flat file, it is compact and efficient to read sequentially, but there is no fast way to access it randomly. This is especially true of variable-length data such as documents, names or strings. To allow for fast random access, most databases store information in a single file using a data structure called a B-Tree. This structure allows for insert, deletion, and search to be fast, but it can use up to 50% more space than the original file. Typically, however, this is not an issue as disk space is cheap and larger, while the primary tasks usually require fast access.
For more information:
http://en.wikipedia.org/wiki/B-tree
Looking carefully into the MySQL docs, we find that indices may be optionally set to "BTREE" or "HASH" type. Inside a single MySQL file, multiple indices are stored which may use either data structure.
Although safety and concurrency are important, these are not WHY databases exist, but added features. The very first databases exist because it is not possible to randomly access a sequential file containing variable length data.