Skip to main content

Wed, Aug 28, 2019 1:11 AM

Answered

Adding more RAM in vSZ for enhancement of performance

Current Specs of the VM (virtual smartzone)



can anyone help me how many users do 1GB of RAM can accommodate? sometimes the vSZ is experiencing "out of service" and the utilization is high. The vSZ is being used in a University most of the users accessing the APs are students.

Thank you.

Responses

129 Messages

 • 

2.4K Points

a year ago

There are multiple resource levels. I don't think you'll gain much by adding a gig or two except maybe for cache performance. The release notes have the different levels that will get you number of clients, etc but there is nos specific gain from having more ram

10 Messages

 • 

182 Points

i understand of having no gain for additional RAM

but how about lack of RAM / Memory for example: if we have 100 users and at least we need 1GB of RAM per 5 users then we need 20GB excluding the overhead of the Ruckus Virtual appliance and we only configured 16GB of RAM

There is one troubleshooting that Ruckus Support did and suggested that there is 20MB of memory remaining for the Virtual Appliance. This is one of the concern we have for the resources being used bu the Virtual appliance

We need to know the RAM allocation per user??

thank you,

129 Messages

 • 

2.4K Points

If you look at the release notes and install guide you'll notice they mentione resource levels. If your VM is not aligned with one for the resource levels it should not come up. The resource levels defines the numbers of clients and aps that can be handling. Check against those guides to see if you're exceeding those numbers.

Also, in your previous post you never mentioned what hypervisor you're running nor what back end storage or how many other VMs are sharing the host.

Iny lab I had run smartzone on a smaller server (while satisfying ram and cores requirements) but performance was awful since my CPU cores were very underpowered.

10 Messages

 • 

182 Points

Hi Diego thanks for the response.
The hypervisor that is being used is SANGFOR Hyper Converged Infrastructure Solution and it runs on HDD, I'm not sure how many VMs are sharing the host. This issue occurs even before it was recently migrated to SANGFOR HCI.
Thanks

129 Messages

 • 

2.4K Points

Not sure what you were using before. But at least out of the box this would not be a supported hypervisor. Do you have any way to see how heavily loaded the base hypervisor is? Like how much CPU / ram / disk io is being used ?

Its likely that HCI is using kvm underneath but from what I remember it's only officially supported on KVM with centos.

You're not using a virtual datapath aren't you?

178 Messages

 • 

2.9K Points

a year ago

Should be using what the release notes show in the required resources tables. If you are still having problems with that level of resources, you could look into a clustered environment. As long as you don't use some features, no additional charge for adding another node (I believe).