I recently had to import some quite large SQL dumps to my local machine for data analysis purposes. There were 10 different .sql dumps in total, and each of them were sized more than 100GB.

Naive Attempt

First of all, I tried to import each by using the < operator:

mysql some_database < some_dump.sql

Unsurprisingly, each task took very long to finish, approximately 6-7 hours per dump file, in a brand new laptop with i7 CPU, 16GB RAM and 1TB SSD. Since I didn’t have 60 hours for importing all, I had to take a look for solutions that can potentially speed up the process.


After Googling and reading through a set of similar issues on Stackoverflow, I came across an answer, referencing to Vadim Tkachenko and explaining why it was taking so long to import.

After grasping the reasons behind, I changed the InnoDB settings (my.ini) as follows:

innodb_buffer_pool_size = 12G # 60% - 70% of your RAM size
innodb_log_buffer_size = 16M # 16M or 32M is fine
innodb_log_file_size = 3G # 25% of buffer pool size
innodb_write_io_threads = 32 # 32 is fine, 64 is maximum
innodb_flush_log_at_trx_commit = 0

That was it after restarting MySQL:

$ sudo service mysql restart


New settings were clearly effective on the import time of dump files. Each import task started to take around 20-30 minutes after fine tuning the settings.