[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: OpenShift HA/Data Tier scaling Questions




----- Original Message -----
> From: "Aaron Knister" <aaron knister gmail com>
> To: users lists openshift redhat com
> Sent: Sunday, December 8, 2013 10:26:21 PM
> Subject: OpenShift HA/Data Tier scaling Questions
> 
> 
> 
> 
> 
> 
> 
> I'm excited to see that "HA features" are coming to OpenShift for the
> application tier. I have some questions about them that I can't find
> the answer to and I'm hoping someone here knows the answers:

When you say "OpenShift", which OpenShift project are you interested in using - Origin, Online, or Enterprise?

> 1. Is integration with external load balancers supported? e.g. BigIP
> F5

Yes - the hooks are there for manual implementation, and I believe someone's working on reference implementations including BigIP. Are you looking at using

> 2. Is there a plugin API for supporting external HA components? e.g.
> BigIP F5, LVS

Have a look at https://github.com/openshift/openshift-pep/blob/master/openshift-pep-005-available-web.md

> 3. What about HA for the database tier, is there support for database
> failover?

That is something we're working on, but not ready for use yet.

> For the data tier, I'm looking to do some serious scaling (150GB Solr
> Database). Are there plans to support different gear sizes for
> different tiers of the application? The web fronts for an
> application don't need 150GB of disk quota or 16GB of ram each. In
> my case they need far, far less than this. For example, I'd like to
> have an application that uses one size of gears for say the tomcat
> gears and a different size for the mongodb gears.
> 
> 
> Also, are there any plans for scaling of databases (i.e. adding read
> replicas)?

Setting different gear sizes within the same application is ready in Origin master, I believe - look for that capability to roll out shortly.

I doubt OpenShift is going to be a good place to locate large, scalable, HA databases for a while yet. For your project it would probably be best to handle the database separately and point to the external DB(s) with environment variable(s).


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]