存储日志的最佳数据类型

Question

I have a MySQL table named requests . It's growing so fast. Currently, it has 3 million rows. Also, the current table engine is "InnoDB".

Few days ago, I got this error:

ERROR 1114 (HY000): The table is full

I've resolved the problem (temporary I guess) by adding this to MySQL configuration:

innodb_data_file_path = ibdata1:10M:autoextend

---

But still sometimes my serves goes to down and I have to restart it to make it available. Recently, I've made that table empty and I got rid of being down.

But it will be full again soon. Any idea how can I fix that issue? Should I use another engine for that table? Should I use compressed type-row for it? or what?

Noted that, sometimes I need to select from that table and it has one index on a column (in addition to pk)

Answer 1

You are limited by disk space. You must keep an eye on that and take action before you get to "table full".

If you store the data in a MyISAM table, you can go twice as long before getting "table full". If you simply write to a disk file (no database, no queries; etc), you can squeeze a little more in before "disk full".

Do you need to "purge" old data? That way you could continue to receive new data without ever hitting "table full". The best way to do that is via InnoDB and PARTITION BY RANGE(TO_DAYS(...)) . If you purge anything over a month old, use daily partitions. For, say, 90 days, use weekly partitions. More discussion: http://mysql.rjweb.org/doc.php/partitionmaint

What will you do with the data? Analyze it? Search it? What SQL queries do you envision? Answer those; there could be other tips.