Many of our MyDBR queries are very heavy and are negatively affecting our production application. I am interested in creating a "Read Replica" at RDS which is essentially a hot copy of the main db to be used for read-heavy things like MyDBR. However, I am not sure how I would configure this since MyDBR appears to be designed to be installed on the same MySQL server as the data. Any ideas?
Using MyDBR with an Amazon RDS Read Replica
(5 posts) (4 voices)-
-
myDBR uses same database connections to both myDBR itself and to the user database, so using read replicas are bit problematic. You could make it work with remote server calls, but that is not really an optimal solution.
Other solution is to use another DB instance where dump/load the production database. Depending on freshness of the data you could adjust the time when to do the loads (read replicas may also be lagging behind).
By far the easiest solution is to optimize the reports to be more efficient. It is not that common that one has so heavy reports that should affect the application use. This of course depends on the reports / database. myDBR itself is quite light on resource usage.
--
myDBR Team -
Hi myDBR Team,
Are there any updates to the above as we do want myDBR database in RDS but for scalabilty/performance reasons, would like to use read replica. Is it something in your roadmap ?
-
There has been quite a bit of development since this old forum thread was created.
myDBR has full support for distributing queries between the master and replicas. You have the flexibility to define myDBR to use replicas based on the report level. This functionality is available in the OEM version.
Alternatively with Premium license, you can set up myDBR to use the replica by adjusting the default value in the user/defaults.php file to:
$mydbr_defaults['db_connection']['disable_writes_to_db'] = true;
--
myDBR Team -
We are using AWS read replicas in our reports and it helps greatly to offload analytical tasks from the writer instance.
Reply
You must log in to post.