Hey Jeff,

Just wondering, have you had a chance to check out our Casandra tutorial 
<https://scalr-wiki.atlassian.net/wiki/display/docs/Creating+a+Cassandra+Role>? 
 
Would the automation/scripts we walk through in that tutorial perhaps help 
to alleviate the need for all cluster member IPs at install and instead 
dynamically reconfigure as nodes join/leave the cluster?  Let us know what 
you think.

Many thanks,
Wm. Marc O'Brien
Scalr Technical Support

On Tuesday, February 14, 2017 at 11:27:50 AM UTC-7, Jeff Gojkovich wrote:
>
> We have an application that requires a few components to be installed in 
> series and some of those components need to be built simultaneously.  First 
> we have cassandra which requires the IPs of all hosts that will be part of 
> the cluster during install time.  We don't have those IPs until 
> the instances are launched which means we need the simultaneous launch of 
> that farm so that we can discover all of the members in that cluster. 
>  After this it will move to the next components which aren't near as picky 
> with other members being up and available.  So if scalr could go through 
> each farm role sequentially and launch the all of the farm role instances 
> simultaneously that would be ideal.
>
> On Monday, February 13, 2017 at 5:48:06 PM UTC-7, Marc O'Brien wrote:
>>
>> Hey Jeff,
>>
>> This functionality is not currently available in Open Source or 
>> Enterprise Scalr.  As of today only the Sequential and Simultaneous launch 
>> options are available.  Keep in mind that even if configured for Sequential 
>> launch Scalr is only able to place the API call with your Cloud Provider 
>> for handling and it is possible that your observed launch order may in some 
>> cases differ from what you have defined in the Farm Designer.  Typically 
>> this is par for the course in the Cloud as failures may occur at any time, 
>> and the right solution here is to alter the way of thinking about 
>> application architecture so that you account for the typical types of 
>> failures that can and will occur for cloud resources. (Ref. Pets vs. Cattle 
>> ideology) Ideally your application and the Orchestration logic you 
>> implement will allow for dynamic (re)configuration of your nodes and 
>> services to maintain operability, regardless of the order launched or 
>> unexpected scale in/out events.  If this is a hard requirement for your use 
>> case you may be able to work around this by using custom scaling metrics or 
>> our API to increase autoscaling minimum from 0 to your desired count for 
>> each subsequent Farm Role in sequence after the successful simultaneous 
>> launch of your first Farm Role, but of course this will require quite a bit 
>> of logic design on your end to achieve your desired results.  Having said 
>> that, we are always looking for feedback and new use cases from our users 
>> so that we can ensure Scalr meets our customer's needs as efficiently as 
>> possible.  If you do have a use case where this functionality would be a 
>> hard requirement, could you please provide any details you may be able to 
>> share about such a use case?
>>
>> Many thanks,
>> Wm. Marc O'Brien
>> Scalr Technical Support
>>
>>
>>
>>
>> On Monday, February 13, 2017 at 5:27:53 PM UTC-7, Jeff Gojkovich wrote:
>>>
>>> Is there a way within Opensource Scalr to launch an entire farm role 
>>> simultaneously and then each farm role serially?
>>>
>>> --
>>> Jeff
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"scalr-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to