added replication support #267
Conversation
| var sequel = new Sequel(schema, sqlOptions); | ||
|
|
||
| // Build a query for the specific query strategy | ||
| try { |
There was a problem hiding this comment.
this seems odd, I may have squished someone elses commit here.
There was a problem hiding this comment.
Try pulling in the latest from master
|
Does this change the connection name dynamically, too? I mean, the logs are useful if they also log which server from the pools you're connecting to? |
|
I wouldn't call this I like the initiative on this PR, 👍 for that. |
| @@ -1117,8 +1106,9 @@ module.exports = (function() { | |||
| * @param {Function} fn | |||
| * @param {[type]} cb | |||
| */ | |||
| function spawnConnection(connectionName, fn, cb) { | |||
| function spawnConnection(queryType, connectionName, fn, cb) { | |||
There was a problem hiding this comment.
Should be careful that this doesn't break anywhere. Did you check for all uses? Does this get exposed to user code? (if so, it's a BC break).
There was a problem hiding this comment.
Looked like an internal call, but someone else may be more aware of the ramifications.
|
@RWOverdijk this naming convention was taken from http://docs.sequelizejs.com/en/1.7.0/docs/usage/#read-replication as for logs yes it logs which index and which pool its from |
|
Ah okay that makes sense. Still though, replication...? :p |
|
I mean it makes sense to me from a DBA perspective because thats how the servers are actually configured for replication. |
|
Exciting stuff! |
|
@icodeforlove @RWOverdijk @devinivy this is interesting. From a glance this actually looks like something we can support on pretty much all adapters by creating multiple connection pools. Although in the sequelize docs it says: Not being a DBA how does this work with transactions? Does it start a transaction on both the read connection and the write connection? |
|
Read replication is a feature of certain types of databases. The databases coordinate with each other to keep multiple read replicas in sync with some master database. Transactions should be supported– I believe they occur on the master database, then the replicas lag slightly behind, catching-up on all the transactions that have occurred on master. |
Tests still need to be made, but heres a basic proof of concept of how we would handle multiple connection pools.