VMware Cloud Community
itbraindead
Contributor
Contributor
Jump to solution

vrops actions

Dears I want to , on pressure moments, increase on the fly the memory/cpu of vms. As far as i know this is not possible on vsphere, but i

discovered that there are two actions on vrops that can fit on that task,  Set Memory for Virtual Machine Action and various set cpu actions.

I want to know if using that actions on a memory/cpu use alarm can "on the fly" add memory or cpu to a vm. Any experience that somebody can share?

it works as expected?

I know that to remove memory or cpu a reboot is needed , and that vrops includes this step on his actions.

thanks for the help.

Greets!

1 Solution

Accepted Solutions
daphnissov
Immortal
Immortal
Jump to solution

I can give you my experience as well as cautions. First, to have that (or any) action triggered automatically, it needs to be item #1 in a symptom. In other words (especially if you roll your own), even if it's in the list of available actions, if it isn't #1 it won't get triggered. This was at least true until 6.6.1. I'm not sure if that's changed in 6.7 or now 7.0. Second, you need to be really careful when using any of these automated actions, but especially the "add X to VM" because vROps doesn't know whether the behavior which is triggering the symptom is legitimate or not. Either way, it's going to add resources. So what I've found with some customers, especially those who don't really watch and observe system behavior very carefully, is there is a see-saw effect going on. vROps adds resources because someone did X and tripped the alarm. Then it shows up on the "oversized VMs" report as having too many resources when that behavior goes away. So you take an outage to remove the resources that were added. You go back and forth which just creates administrative overhead and headaches. So, lesson is to use with caution and don't just flip the switch on and walk away.

View solution in original post

5 Replies
virtualg_uk
Leadership
Leadership
Jump to solution

Hello and welcome to the forums!

In vSphere (Depending on your license) you are able to hot-add memory and CPU to virtual machines (As long as the VM OS supports this)

vROps, can make use of this feature in vSphere by it's actions.

As a basic example, if enabled, if a CPU alarm triggers on a VM then you can apply an action to add more CPU to the VM. This can be done without a reboot if hot-add is enabled on the VM and is supported by the VM OS. If hot-add is not supported, then a VM reboot will be required and within vROps actions you can permit this to automatically occur.

I hope this helps?


Graham | User Moderator | https://virtualg.uk
Reply
0 Kudos
itbraindead
Contributor
Contributor
Jump to solution

thanks graham, just looking for some experience on using this feature , i will be glad to hear some feedback if someone used it in the past or now. and if it works as expected or there are some tricks to make this working.

Reply
0 Kudos
daphnissov
Immortal
Immortal
Jump to solution

I can give you my experience as well as cautions. First, to have that (or any) action triggered automatically, it needs to be item #1 in a symptom. In other words (especially if you roll your own), even if it's in the list of available actions, if it isn't #1 it won't get triggered. This was at least true until 6.6.1. I'm not sure if that's changed in 6.7 or now 7.0. Second, you need to be really careful when using any of these automated actions, but especially the "add X to VM" because vROps doesn't know whether the behavior which is triggering the symptom is legitimate or not. Either way, it's going to add resources. So what I've found with some customers, especially those who don't really watch and observe system behavior very carefully, is there is a see-saw effect going on. vROps adds resources because someone did X and tripped the alarm. Then it shows up on the "oversized VMs" report as having too many resources when that behavior goes away. So you take an outage to remove the resources that were added. You go back and forth which just creates administrative overhead and headaches. So, lesson is to use with caution and don't just flip the switch on and walk away.

itbraindead
Contributor
Contributor
Jump to solution

Thanks for the feedback daphnissov, do you have any example of a symptom that can be used to increase memory when vm is in need of that resource?

contention/used/requested?

thanks!

Reply
0 Kudos
daphnissov
Immortal
Immortal
Jump to solution

You can go and look at the symptom list yourself and then filter by those actions. There are a few that can trigger this.

Reply
0 Kudos