[HyperDB] Replication Lag Detection

Brian Layman bulk at thecodecave.com
Wed Jul 20 16:42:50 UTC 2011

On 7/20/2011 12:18 PM, Barry Abrahamson wrote:
> No real object caching - we are only using the cache to store the 
> states (lagged or not lagged) for the various servers because we don't 
> want to query the lagged state before every query.
Ah, ok I was confused by "In case replication breaks on all slaves, we 
would rather return old data then overload the master with read queries 
and cause an outage." and I thought that meant it returned cached data, 
but you simply mean you then query an existing slave even though it is 
lagged, right?

> Someone else needs to use it successfully in production before we make a point release :)
I'll see what I can do.  One client will likely say no, another might 
say maybe.


*Brian Layman *

Founder and Developer
brian at ehermitsInc.com <mailto:brian at ehermitsInc.com>
twitter: @brianlayman

*eHermits Inc.*

http://eHermitsInc.com <http://ehermitsinc.com/>


Cell: 330-887-2283

More information about the HyperDB mailing list