I looked up the mysql_replication_adapter - interesting indeed.
Which scheme are you using to select a slave (random or the block syntax ?)
I'm also curious as to what happens to queries that are part of a
transaction that involves an update ? In the PHP app, we do the reads
from the master database so as to ensure that the transaction will stay
consistent. I don't see this documented in the plugin - any idea how it
works ?
Shanti
David Schleimer (JIRA) wrote:
[
https://issues.apache.org/jira/browse/OLIO-55?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
David Schleimer updated OLIO-55:
--------------------------------
Attachment: mysql_replication.patch
Patch to add ability to use multiple mysql databases in a master-slave
configuration. This is a modified version of the adapter mentioned in the
initial description for this issue. Specifically it was ported to rails 2.2.2
(available at
http://github.com/findchris/mysql_replication_adapter/tree/master), and then
backported to rails 2.1.2 and modified to default to using the slaves where
possible.
add support for mysql master-slave clusters
-------------------------------------------
Key: OLIO-55
URL: https://issues.apache.org/jira/browse/OLIO-55
Project: Olio
Issue Type: Improvement
Components: rails-app
Environment: N/A
Reporter: David Schleimer
Assignee: Shanti Subramanyam
Attachments: mysql_replication.patch
Original Estimate: 96h
Remaining Estimate: 96h
Add support to the rails app for reading from multiple separate mysql hosts
while writing to a single mysql master. Specifically using a modified version
of mysql_replication_adapter. the original can be found at
http://rubyforge.org/projects/mysql-replicate/ The specific modifications are
to default to using the slaves instead of requiring that you explicitly mark a
query as to be directed to a slave.
