Ask Your Question

rpga95's profile - activity

2015-12-14 08:41:21 -0500 received badge  Famous Question (source)
2015-12-14 08:39:47 -0500 commented question heat stack instance names

I'd like to know about this too..

2015-08-14 06:43:54 -0500 received badge  Notable Question (source)
2015-07-30 13:35:46 -0500 commented question The X-Auth-token created using REST api call doesn't work? Whereas the token generated using commandline works.

I believe port 35357 corresponds to the admin token in the first few lines of "/etc/keystone/keystone.conf" To use the token you generated, use keystone at port 5000.

~RP

2015-07-29 12:44:30 -0500 received badge  Famous Question (source)
2015-07-29 12:41:29 -0500 received badge  Nice Question (source)
2015-07-28 14:42:26 -0500 commented answer How can Heat gracefully shutdown a server when scaling down?

Thanks so much for this! Exactly what I was looking for. I am actually autoscaling a custom nova server template within my template for the autoscalinggroup.

~RP

2015-07-28 14:39:13 -0500 commented answer Sending Heat Scale URLs to Instances

That's definitely a more secure and easier to manage option. Thanks for your help.

~RP

2015-07-28 14:39:13 -0500 received badge  Commentator
2015-07-28 14:37:50 -0500 received badge  Scholar (source)
2015-07-28 14:36:15 -0500 received badge  Popular Question (source)
2015-07-28 13:48:45 -0500 received badge  Notable Question (source)
2015-07-28 13:48:45 -0500 received badge  Famous Question (source)
2015-07-28 13:48:45 -0500 received badge  Popular Question (source)
2015-07-28 12:59:14 -0500 commented answer when ever i create alarm is state is insufficient data

In my experience, you need to set the reporting times in "/etc/ceilometer/pipeline.yaml" to be less than the period when you create the alarms.

~RP

2015-07-28 12:43:52 -0500 commented answer Unable to launch stack in kilo heat

I would start by checking the heat logs. "/var/log/heat/*"

~RP

2015-07-28 12:34:01 -0500 received badge  Editor (source)
2015-07-28 12:32:14 -0500 received badge  Notable Question (source)
2015-07-28 12:32:14 -0500 received badge  Popular Question (source)
2015-07-28 12:20:34 -0500 commented question my instance creation fails and logs reveals ""No valid host was found. There are not enough hosts available"

I've solved the same error by restarting my compute.

~RP

2015-07-28 12:18:35 -0500 received badge  Famous Question (source)
2015-07-28 12:18:35 -0500 received badge  Notable Question (source)
2015-07-28 12:18:35 -0500 received badge  Popular Question (source)
2015-07-27 09:57:37 -0500 asked a question Heat Personality Limit

Hi,

I am using file injection with Heat, and I was wondering why the limit is only 5 files. Is there a particular reason for this and is there a way to change it? I looked in the heat engine's server.py and found how the error message is generated and that the value is stored in a list. Is it safe to change the value?

~RP

2015-07-24 09:39:38 -0500 commented answer Heat Downscaling -- Graceful Delete

Thanks, @Steve Baker. Is there a way to change the Heat Downscale policy to choose a different VM, more intelligently?

~RP

2015-07-23 13:19:49 -0500 asked a question How can Heat gracefully shutdown a server when scaling down?

Hi,

Does anyone know if there are any advanced parameters for Heat Autoscaling, particularly with Downscaling? For example, how to have certain scripts run at shutdown or even to shutdown gracefully rather than hard termination?

Thanks,

~RP

2015-07-23 11:17:20 -0500 commented answer Can we assign instance names to autoscaling servers?

I'd like to know the same thing, but thanks for the insight @don!

~RP

2015-07-22 08:30:04 -0500 commented question Ceilometer change state

Yes, I'd go through the ceilometer logs in "/var/log/ceilometer/.." since they usually offer some insight as to what's going on.

~RP

2015-07-20 13:16:48 -0500 asked a question Sending Heat Scale URLs to Instances

Hello,

I have set up Heat with the autoscaling capability, and I'm wondering how I would go about sending the scaleup and scaledown URLs to the instances for inside-the-os alarming. In order to create the autoscalinggroup, you must have all parameters for the instance template. In order to have all parameters, you must have the scale URL webhooks to pass to the instance template. And in order to make the scalingpolicies, you must have the autoscalinggroup ID. I don't see a way to do this via the template without having circular dependencies. Any thoughts?

~RP

UPDATE:

I've found a way around this apparent issue. I created some scripts to go to the REST API and pull the URLs from my stack via the output fields. This would be a script injected to the instances by Heat and run continuously. Here's my next problem, if you're following this: Personality Limit for File Injection by Heat

2015-07-17 08:54:13 -0500 received badge  Enthusiast
2015-07-15 14:10:07 -0500 commented answer How do you setup autoscaling based on memory?

@ichi-the-one: I have set up autoscaling based on cpu utilization, but have you made any progress with setting up autoscaling based on memory?

2015-07-15 13:41:50 -0500 asked a question Handling Heat Autoscale with multiple alarms

Hi,

I have created a Heat autoscaling stack of resources using a template which uses ceilometer alarms, an autoscaling group, and scaling policies. This works perfectly for high and low cpu_util alarms since the alarms trigger a webhook that handles the scaling via the scaling policies. Is there a way in Heat to make a scaling decision based on multiple alarms going off at the same time? For example, if cpu_util alarms high and memory alarms low, which decision would be taken? Currently, the alarm which triggers first of the two occurs first. This is an issue because if cpu_util was high and scaled the stack up, then it might be "ok" and memory is still low. After the cooldown period, the stack would scale down by one. After that, the cpu_util may scale back up, and this cycle would continue. All I'm really wondering is if instead of directly triggering a webhook, is there a way to have the alarms compared against one another?

I think a potential way to handle high/low situations is scale up if any alarm is high and only scale down if all alarms are low.

Thanks in advance for any help,

~RP

2015-07-08 10:21:04 -0500 received badge  Student (source)
2015-06-26 09:42:57 -0500 received badge  Famous Question (source)
2015-06-25 09:58:25 -0500 received badge  Organizer (source)
2015-06-25 09:48:14 -0500 received badge  Notable Question (source)
2015-06-25 09:48:14 -0500 received badge  Popular Question (source)
2015-06-22 09:40:32 -0500 commented question Kilo(OSP7) Deployment failing on NoValidHost: No valid host was found.

Hello, I ran into this very same problem. If you have enough physical resources to support the stack of resources, make sure the network you are using for the stack is NOT external. I tried deploying a heat stack on an external network and that was my problem. Hope this helps.

~RP

2015-06-22 09:00:23 -0500 asked a question Heat Downscaling -- Graceful Delete

Hello,

I have autoscaling working with Heat, but I need to know how the downscale chooses a resource (in my case, a VM) to delete or if there is a graceful method in place already. I'd like a way to make sure that connections are not dropped as a result of the downscale. Thanks.

~RP

2015-06-22 08:32:58 -0500 received badge  Supporter (source)