Wednesday, March 28. 2012MySQL 5.6.4-m7 is made of pure winTrackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Great to see this! Looks like we (the MySQL family together) continue to tackle the bottlenecks one by one...
When you say "internal stall protection", is that something in InnoDB or in your app? Can you elaborate on the back-of-napkin calculation you use to calculate redo log size?
The application in question monitors the query response time at the database wrapper class level and backs off when the response times are getting out of bounds. The backoff is essentially wait-time-squared with a maximum cap and a few corrective factors.
The back-of-napkin calculation looked at the amount of dirty pages desired (maximum would be innodb-buffer-pool scaled with max-dirty-pct) and the ratio of actual dirty pages observed in relation to actual redo log usage observed. In the application and for this type of workload, that seemed to be relatively stable at a ratio of 1:6 to 1:8. So in our environment we wanted something in the range of 192 GB RAM (essentially all of it being buffer pool) divided by a ratio of 6 (or 8) = 32G (or 24G) of redo log. The actual value does not really matter, as the ratio is not overly stable, but we wanted to be able to hit the right order of magnitude and get the first digit right. There is little value in tuning this too much: Running at 100% efficiency just means you have zero elasticity. |
QuicksearchComments
about Wed, 31.10.2012 13:46
My friendly Offtopic Channel
in IRC has horror stories
about this, it is Halloween,
after all.
"Well, I do
have [...]
about Mon, 15.10.2012 09:41
What if we use SSDs. Can we
then assume memory limits
are not as significant
because SSDs can handle
random access [...]
about Fri, 05.10.2012 07:46
In fact, I am pretty certain
that this also breaks
mysql_upgrade or has the
potential to do so.
about Fri, 05.10.2012 06:44
It is quite a bit worse than
that: Right now, the default
RPM as distributed by Oracle
will fail to install if it
[...]
about Thu, 04.10.2012 18:52
> MyISAM in mysql.* has to go
before GTID can be
sucessfully deployed.
I
think you meant to say
before it can b [...]
about Thu, 04.10.2012 18:36
Sorry, it wasn't clear given
that none of this is
actually mentioned.
about Thu, 04.10.2012 13:47
You can turn it on with FLUSH
PRIVILEGES once the tables
exist.
about Thu, 04.10.2012 11:34
I think I have to install a
comment voting plugin just
so that i can +1 what Justin
just wrote.
Categories
Blog AdministrationDisclaimerAll examples, code-snippets, advice, trivia, etc. are supplied in the hope that they are useful, but no warranty of any kind is given. Use of trademarks is not intended as a challenge to them.
ImpressumCopyright 2005-2012 bei den jeweiligen Autorinnen der Beiträge. Die Infrastruktur von MySQL-dump wird betrieben von:
Azundris, Immanuelkirchstraße 18, 10405 Berlin, Germany
|