Reputation
14,103
Top tag
Next privilege 15,000 Rep.
Protect questions
Badges
2 20 43
Newest
 Talkative
Impact
~1.1m people reached

Mar
24
comment MySQL Too Many Connections (Max Connections Limit Not Reached)
The most useful thing to do, first, is to capture the exact wording of the error, including error number and sqlstate code if possible. Also, how exactly are you finding the "highest connection count?"
Mar
22
awarded  mysql
Mar
21
comment Give third-party access to AWS hosted MySQL
If this is RDS then give 'em access to a read replica. They not only can't modify your database, they also can't impair its performance with badly written queries, which is always a risk.
Mar
20
comment How to verify galera is synced
@jerichorivera it should equal the number of nodes, but if it isn't, that does not necessarily mean this node has a problem. With 3 nodes, if "node x" suddenly returns a value of "2" here, then "node x" would actually be one of the good nodes... So, while useful information, it doesn't really give you any new information about this node.
Mar
20
comment How to verify galera is synced
@Som3guy yes, 'Synced' is correct.
Mar
20
answered How to verify galera is synced
Mar
19
comment Select query on an index column with gap locking
A SELECT query run in isolation wouldn't need to lock the gap, but you're building a derived table with the select and using it to join into the update. That's a different scenario entirely.
Mar
19
comment MySQL Server won't start
This log isn't useful. You need to find the real one, possibly at /var/log/mysql/$(hostname).err...
Mar
19
comment Setting up a redis or memcached slave table with a MySQL (or other traditional database) as the master
InnoDB tries to work in memory. Of course all changes have to be written to disk immediately, otherwise you'd have no ACID but yes, the InnoDB buffer pool is an underappreciated aspect of InnoDB.
Mar
19
comment Setting up a redis or memcached slave table with a MySQL (or other traditional database) as the master
It really depends on why you are wanting this -- what specific problem does it solve? It's usually best to ask about the problem rather than about your attempted solution. InnoDB in MySQL is all-memory-based, disk-backed, given a system with sufficient memory for the data and correct provisioning... and InnoDB can be queried with a memcache client, for pretty blazing speed, since the SQL parser is bypassed when this feature is enabled and used. dev.mysql.com/doc/refman/5.6/en/innodb-memcached-setup.html
Mar
7
awarded  Enlightened
Mar
7
awarded  Nice Answer
Mar
5
comment Preventing MySQL index corruption
ANALYZE TABLE would probably also "fix" it. This doesn't look like "corruption," merely inaccurate index statistics.
Mar
4
comment How to observe output from query in Workbench 6.3?
Running a "script" is assumed to either succeed, or fail with an error. Output from scripts is discarded.
Mar
4
answered Hybrid Synchronous / Asynchronous Replication?
Mar
4
comment mysqldump not consistent after stopping slave
If left the slave stopped while the entire mysqldump ran and you got the coordinates correct, then what you describe should be impossible.
Mar
2
comment SQL Timeout when updating specific rows that goes away when other columns are updated
I would suggest dropping it down to something like 32M and see if that resolves the issue. When a table that has contributed to the query cache is modified, all of the cached queries linked to that table are purged, whether they need to be or not. It's a very useful feature, which I usually enable, but it doesn't scale up very well to large sizes. I'm not sure when the lock is acquired -- the query cache is actually checked before the query is parsed, because the cache operates against the raw bytes of the query, and only if the query starts with 'SEL'.
Mar
2
comment Mysql hung at 41% CPU Usage
Well, one thing you can rule out for certain is Slave has read all relay log; waiting for the slave I/O thread to update it. The slave SQL thread has read up through the end of the last relay log and has absolutely no pending work to do; it will have nothing to do until the slave IO thread receives the next event from the master and writes it to the relay log, so it's waiting -- and idle. Your config change didn't do anything here. Waiting for master to send event means the IO thread, in turn, is fully caught up with the master, and is also idle, so nothing to see here, either.
Mar
2
comment SQL Timeout when updating specific rows that goes away when other columns are updated
Waiting for query cache lock is a sign of a potentially too-large query cache; you have threads blocking on a global mutex, here. What's the result from SELECT @@QUERY_CACHE_SIZE;?
Feb
24
comment How do I efficiently partition a large MySQL (Aurora) table with multiple search columns?
You're inserting 432 million rows per day? (almost 13 billion rows per month?!)