Posted by: Peter Brawley
Date: June 27, 2017 12:54PM

You do need to fix the slow queries.

You're running the binary log and the general log. Why do you need the latter? Everything's being written four times.

Have you confirmed that the crashes occur during the bulk inserts?

The load on InnoDB looks very heavy. It's doing (Innodb_buffer_pool_reads + Innodb_buffer_pool_pages_flushed) / Uptime_since_flush_status = 672 buffer pool reads/sec, writing over 6G/hr to its logfiles, 5 log fsyncs/sec.

It's often recommended that InnoDB logs be configured to hold about an hour's data ( Yours are a third of that.

For starters, have you considered setting innodb_log_group_home_dir to another HD?

Options: ReplyQuote

Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.