[openstack-dev] [OpenStack][Heat] AutoScaling scale down issue
cbjchen at linux.vnet.ibm.com
Thu Dec 12 12:10:55 UTC 2013
On 12/11/2013 11:43 PM, Jay Lau wrote:
> Here come a question related to heat auto scale down.
> The scenario is as following:
> I was trying to deploy hadoop cluster with heat Auto Scaling template.
> When scale up a slave node, I can use user-data to do some post work
> for configuration file on hadoop master node base on the information
> of slave node (The mainly configuration file is conf/slaves as I need
> to put slave node to this file);
> But when scale down, seems I have no chance to do some configuration
> for the master node (Remove the scale down node from conf/slaves) as
> master node do not know which slave node was scale down.
Yes. There isn't a way to add a hook to do any sort of "post work" after
an instance is brought down. That is why AutoScalingGroup and
LoadBalancer resource usually come in pair, even though
LoadBalancerNames isn't a mandatory property. Internally
AutoScalingGroup triggers a load balancer reload each time it's resized,
but that doesn't happen to any other resources (like a customized
instance in your case). This might be something we can improve in the
> Does anyone has some experience on this?
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the OpenStack-dev