Web lists-archives.com

Re: threads in cleaning up mode






On 4/2/2016 10:36 AM, geetanjali mehra wrote:
Hi to all,

For many days, I am struggling with the problem of  increasing history list
length on mysql 5.6.29. Application that this server is running IOT based.
Also, this server has so many threads running in sleeping mode. SHOW ENGINE
INNODB STATUS shows all these threads in *cleaning up* mode.
I tried all the options to reduce history list length. But it is constantly
increasing.
Below are the current settings of purge related threads:

innodb_max_purge_lag       | 1000000 |
| innodb_max_purge_lag_delay | 0       |
| innodb_purge_batch_size    | 1       |
| innodb_purge_threads       | 8       |

Also, please let me know that whether cleaning up mode of threads and
history list length are correlated.

Anty help?


Best Regards,
Geetanjali Mehra
Senior Database Administrator


No, the "cleaning up" status and your ever-increasing history are not related. The "cleaning up" status is a very low-impact bug in our code. We simply forgot to reset a flag at the end of that section of the code. It is not indicative of any ongoing operations.

What has happened is one of those idle threads (the sleeping ones) has opened a transaction but never committed it. That is why your history list is never shrinking. As was already said in another response, this is an indication of poor client-side transaction management. If you start a transaction, you must always terminate it or situations like yours will happen. (This is also one of the dangers if someone changes autocommit=0 and forgets to change it back.)


--
Shawn Green
MySQL Senior Principal Technical Support Engineer
Oracle USA, Inc. - Integrated Cloud Applications & Platform Services
Office: Blountville, TN

Become certified in MySQL! Visit https://www.mysql.com/certification/ for details.

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:    http://lists.mysql.com/mysql