Re: mysql replication issue (Handling write load on slaves)?
Posted by:
Rick James
Date: March 16, 2009 10:17PM
Possible reason 1: The slaves are wimpier than the master.
Possible reason 2: Replication is _serial_. The queries (INSERT, UPDATE, etc) are running nicely in parallel on the Master, but the replicated statements are not run in parallel on the Slaves.
Possible reason 3: Slaves are not well tuned.
Possible reason 4: You have big batches of queries that replicate all at once, but could be spread out more during the day.
Possible reason 5: You are simply doing more than can be handled.
Possible reason 6: You need to shard.
To assist further, please provide this for the likely tables:
* SHOW CREATE TABLE tbl\G
* SHOW TABLE STATUS LIKE 'tbl'\G
and also
* SHOW VARIABLES LIKE 'innodb%';
* SHOW VARIABLES LIKE '%size';
* How much RAM is on each machine?
* Show us some of the slowlog entries for INSERTs/UPDATEs for the Master (they won't show on Slaves)
and surround them with [ code ] and [ / code ]
Subject
Views
Written By
Posted
3955
March 15, 2009 08:51PM
Re: mysql replication issue (Handling write load on slaves)?
2097
March 16, 2009 10:17PM
2081
March 20, 2009 02:46AM
2317
March 20, 2009 10:34PM
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.