Error: Tablespace for table xxx exists. Please DISCARD the tablespace before IMPORT

I am fairly new to MySQL and I am getting a pretty interesting error on which I cannot find any help via google and the stackoverflow search.

I am running a local server of MySQL 5.6.10 on MacOS 10.8.3 and manage my database via Navicat essentials for MySQL.

The error I get is that after running and managing my database just fine for a couple of days/weeks something triggers to (it appears incompletely) delete some of the tables I created using queries from within Navicat.

When I try to run queries using these tables, Navicat then warns me that the particular table does not exist. So far so good - here comes the good part:

When I try to CREATE the table, e.g. named "temp", that was previously there, I get the following error message:

Error : Tablespace for table '`database`.`temp`' exists. Please DISCARD the tablespace before IMPORT.

However, if I try to drop the table, or try to discard the tablespace for this table, using

DROP TABLE temp;
ALTER TABLE temp DISCARD TABLESPACE;

I get the following error messages:

Error : Unknown table 'database.temp'
Error : Table 'database.temp' doesn't exist

So that means that I am advised to discard the table space but when I try to do so the table does not exist. Is it possible that there is some type of remnant of this table at a different place where the DISCARD query isn't checking? And does anybody have an idea what could trigger all that - completely randomly as it seems?

As I said, I'm new to the subject and pretty much clueless. I suspect that re-booting my laptop, i.e. resetting my local MySQL server, or maybe user permission rights might have to do with it, but I am just hypothesizing here.

A little late here but generally I've seen this problem occur when you get a 'tablespace full' error when running in a 'innodb_file_per_table' mode. Without going into too much detail (more here), the database server's tablespace is defined by the innodb_data_file_path setting and by default is rather small. Even made larger, the 'tablespace full' can still occur with larger queries and such (lots of non-table 'stuff' is stored in there, undo logs, caches, etc...).

Anyways, I found that if you look in the OS directory where the files-per-table are stored, /var/lib/mysql by default on OSX, /usr/local/var/mysql with homebrew iirc, you'll find an orphaned tablename.ibd file without it's normal companion tablename.frm file. If you move that .ibd file to a safe temporary location (just to be safe) that should fix the problem.

$ ls /var/lib/mysql

table1.frm
table1.idb
table2.frm
table2.idb
table3.idb <- problem table, no table3.frm
table4.frm
table4.idb

$ mkdir /tmp/mysql_orphans
$ mv /var/lib/mysql/table3.ibd /tmp/mysql_orphans/

One caveat though, make sure what ever is causing the problem originally, e.g. long running query, locked table, etc... has been cleared. Otherwise you just end up with another orphaned .ibd file when you try a second time.

Xampp and Mamp Users

Had the same error while importing a database (after emptying it) trough MySQL. I found that i had a tablename.ibd file left while all others were deleted. I deleted it manually from mysql/data/database_name and the error was gone.

If you get the .idb recreated again after you delete it, then read this answer.

This how it worked with me. I had the .idb file without it's corresponding .frm and whenever I delete the .idb file, the database recreate it again. and I found the solution in one line in MySQL documentation (Tablespace Does Not Exist part)

>>1- Create a matching .frm file in some other database directory and copy it to the database directory where the orphan table is located.

>>2- Issue DROP TABLE for the original table. That should successfully drop the table and InnoDB should print a warning to the error log that the .ibd file was missing.

I copied another table .frm file and name it like my missing table, then make a normal drop table query and voila, it worked and the table is dropped normally!

my system is XAMPP on windows MariaDB v 10.1.8

For WAMP [Windows 7 Ultimate x64-bit] Users:

I agree with what DangerDave said and so I'm making an answer available for WAMP Users.

>Note: First of all, you have to go to your ..\WAMP\Bin\MySQL\MySQL[Your MySQL Version]\Data folder.

Now, you'll see folders of all your databases

• Double-click the folder of the database which has the offending table to open it
• There shouldn't be a file [Your offending MySQL table name].frm, instead there should be a file [Your offending MySQL table name].ibd
• Delete the [Your offending MySQL table name].ibd
• Then, delete it from the Recycle Bin too
• Then run your MySQL query on the database and you're done

In my case:

First remove tableName.ibd in your database directory from Mysql and second run:

ALTER TABLE tableName DISCARD TABLESPACE;
DROP TABLE tableName;

In my case the only work solution was:

1. CREATE TABLE bad_table ENGINE=MyISAM ...
2. rm bad_table.ibd
3. DROP TABLE bad_table

This is exactly what i did in mariadb 10.2.16 on fedora when i had a table that showed exactly the same errors in the log file i suppose...

2018-07-11  9:43:58 140323764213504 [Note] InnoDB: The file './database_name/innodb_table.ibd' already exists though the corresponding table did not exist in the InnoDB data dictionary. You can resolve the problem by removing the file.
2018-07-11  9:44:29 140323764213504 [Warning] InnoDB: Tablespace 'database_name/innodb_table' exists in the cache with id 2836 != 2918

your mileage and errors may vary but the main one i assume is

...already exists though the corresponding table did not exist in the InnoDB data dictionary...

with drop table not working as well as alter table...

MariaDB [database_name]> drop table innodb_table;
ERROR 1051 (42S02): Unknown table 'database_name.innodb_table'

MariaDB [database_name]> alter table innodb_table discard tablespace;
ERROR 1146 (42S02): Table 'database_name.innodb_table' doesn't exist

create table also fails like so:

MariaDB [database_name]> create table  innodb_table(`id` int(10) unsigned NOT NULL);
ERROR 1813 (HY000): Tablespace for table '`database_name`.`innodb_table`' exists. Please DISCARD the tablespace before IMPORT

in order to fix this, what i did was first

create table  innodb_table2(`id` int(10) unsigned NOT NULL);
Query OK, 0 rows affected (0.07 sec)

then in the /var/lib/mysql/database_name directory i did the following as root acknowledging the overwriting of innodb_table.ibd causing us issues

cp -a innodb_table2.frm innodb_table.frm
cp -a innodb_table2.ibd innodb_table.ibd
systemctl restart mariadb

then back in the mysql console i issued a successful drop command on both tables

MariaDB [database_name]> drop table innodb_table;
ERROR 2006 (HY000): MySQL server has gone away
No connection. Trying to reconnect...
Connection id:    8
Current database: database_name

Query OK, 0 rows affected (0.08 sec)

MariaDB [database_name]> drop table innodb_table2;
Query OK, 0 rows affected (0.25 sec)

and everything is now all square and i can recreate the one single table...

MariaDB [database_name]> create table  innodb_table (`id` int(10) unsigned NOT NULL);
Query OK, 0 rows affected (0.08 sec)

> EDIT: I was going to add in a > > restorecon -Rv /var/lib/mysql/database_name > > command after the copying of the database to get all selinux contexts > the way they should be, even though we are deleting them from the > database almost immediately, but in the alternative you could just add > the --archive or -a option to the two cp commands, so yes actually the > archive option shortens this: >
> cp innodb_table2.frm innodb_table.frm > cp innodb_table2.ibd innodb_table.ibd > chown mysql:mysql innodb_table.frm innodb_table.ibd > chmod 660 innodb_table.frm innodb_table.ibd > restorecon -Rv /var/lib/mysql/database_name > systemctl restart mariadb > > to just the following which i think is better and it keeps the selinux > context that is set for the already made table. > > cp -a innodb_table2.frm innodb_table.frm > cp -a innodb_table2.ibd innodb_table.ibd > systemctl restart mariadb > > I have replaced the above longer list of commands for the shorter list > which could be shortened still with an *

Deleting/Moving tablename.ibd sure did not work for me.

How I solved it

Since I was going to delete the corrupted and non existing table, I took a backup of the other tables by going to phpmyadmin->database->export->selected tables to backup->export(as .sql).

After that I selected the database icon next to database name and then dropped it. Created a new database. Select your new database->import-> Select the file you downloaded earlier->click import. Now I have my old working tables and have the corrupted table deleted. Now I just create the table that was throwing the error.

Likely I had an earlier backup of the corrupted table.

Solution

However, the easier option is this: restart MySQL, then do the same four steps as follows:

1) created a dummy table in the database;
2) discarded its tablespace;
3) moved the .ibd file into the database folder on the system;
4) attached the tablespace back to the table

This way, the tablespace id on the data dictionary and the file matched; thus importing the tablespace succeeded.

This can give you greater confidence in dealing with some of the InnoDB "gotcha's" during the recovery process or even file transfers.

ref

I got the same error running it on wampserver while trying to create a users table. I found a users.ibd file and after I deleted this file, I ran the migrate command again and it worked. The file on my windows machine was located in wamp/bin/mysql/mysql5.6.12/data/myproject.

Had this issue several times. If you have a large DB and want to try avoiding backup/restore (with added missing table), try few times back and forth:

DROP TABLE my_table;

ALTER TABLE my_table DISCARD TABLESPACE;

-and-

rm my_table.ibd (orphan w/o corresponding my_table.frm) located in /var/lib/mysql/my_db/ directory

-and then-

CREATE TABLE IF NOT EXISTS my_table (...)

Here is the solution steps:

1. backup your database (structure with drop option and data)
2. stop mysql engine service
3. remove the database directory manually from inside mysql/data
4. start mysql engine
5. create new database with any name different from your corrupted database
6. create single table with the name of the corrupted table inside the new database (this is the secret). and it is better to create the table with exactly the same structure.
7. rename the database to the old corrupted database
8. restore your backup and your table will be working fine.

Had exactly the same problem; I'd brew added [email protected] (after previously having 5.5).

The brew defaults for 5.6 are innodb_file_per_table=1 whereas in 5.5 they're innodb_file_per_table=0.

Your existing ibdata1 file (the combined innodb data) will still have references to the tables you're trying to create/drop. Either change innodb_file_per_table back to 0, or delete the ibdata1 data file (this will lose you all your data, so make sure you mysqldump it first or already have an .sql dump).

The other brew [email protected] default that bit me was the lack of a port, so networking was defaulting to unix sockets, and the mysql client kept reporting:

ERROR 2013 (HY000): Lost connection to MySQL server at 'sending authentication information', system error: 32

I added <string>--port=3306</string> to the .plist array, but you could also specify port=3306 in your my.cnf

Run brew services stop [email protected] make your changes then brew services start [email protected]

This error occurs when you suspend some functions. Like running the query below with incorrect foreign key.

set foreign_key_checks=0

I had the same issue. I renamed the database name & import it. Then it works.

Make it simple: Just Rename the Database name. That's it.

Problem is the existing database name has some entries in your folder. That's the problem. Either you need to remove all those related entries or Just Make it simple to rename the database

Trying to drop the tablespace can give you other errors. For me, I got the following error:

DROP TABLESPACE `tablename`

Error Code: 1478. Table storage engine 'InnoDB' does not support the create option 'TABLESPACE or LOGFILE GROUP' 

My solution was to drop the database. This will remove any tablespaces related to it and allow you to create the tables again.

If you have a another server with a good version of the same table you can make a copy(table_copy), transfer the table_copy to the problem server. Then delete the problem table and rename table_copy to table.

For me it helped just go to MYSQL DATA directory under /var/lib/mysql/{db_name} (linux) and drop {table_name}.ibd file which was the same as folder name.

I only delete my old DB located in my localhost directly from wamp, Stop all services, Go to wamp/bin/mysql/mysql[version]/data and I found the DB with problemas, I delete it and start again wamp all services, create again your database and it is done, Now you can import your tables,

The way that I found to "solve" this problem is fairly annoying, but there is a script that handles it.

Essentially, you need the ibdata1 and ib_logfile* files to go away (they contain the mappings of foreign keys, among other things). The only safe way to do this is to export all of your databases, stop mysql, remove the files, start mysql, and then import the files.

The script that helps solve this problem is https://github.com/uberhacker/shrink-ibdata1, even though the stated purpose of this script is different, it does solve the problem.

You can run the following query as a mysql root user

drop tablespace `tableName`

The only way it worked for me was:

1. Create a similar table
2. Copy the .frm and .idb files of the new similar table to the name of the corrupt table.
3. Fix permissions
4. Restart MariaDB
5. Drop the corrupt table

Thank you #DangerDave this solved my problem on Magento 2, and this is how I did

I am on a vps

root@myvps [~]# cd /var/lib/mysql/mydatabasename/

root@myvps [~]# ls

check for tables with no .frm fie (only .idb) and delete them,

rm customer_grid_flat.ibd

System will regenerate tables after running index:reindex command

In case of Homebrew, the directory for data files is /usr/local/var/mysql

to see which my.cnf file is used here are the search locations from my environment /etc/my.cnf /etc/mysql/my.cnf /usr/local/etc/my.cnf ~/.my.cnf

So to clean up this error I did the following

mysqladmin -u root shutdown
rm /usr/local/var/mysql/<dbname>/problemtablename.ibd

Note: in my case i didnt care about the data as it is dev setup, probably happened because i restored my laptop with timemachine

if you have this problem and you don't have another option change the engine to any other engine like 'myisam', then try to create the table.

disclaimer: it is not the valid answer as you may have foreign key constraints which will not be supported by another storage engine. Every storage engine has their own specialty to store and access data, these points also to be taken in account.

Please DISCARD the tablespace before IMPORT

I got same issue solution is below

1. First you have to drop your database name. if your database is not deleting you have flow me. For Windows system your directory will be C:/xampp/mysql/data/yourdabasefolder remove "yourdabasefolder"

2. Again you have to create new database and import your old sql file. It will be work

Thanks

I had to locate my MySQL data directory:

SHOW VARIABLES WHERE Variable_Name LIKE "%dir"

Then force remove that database:

sudo rm -rf