Thanks a lot Rajani & Ilya for your response!

Regarding the CS behavior with saving a copy of the templates on each
primary storage (In my case - VMware datastores), I just want to make sure
I got it right. Lets say I have 50 templates in my environment and 4
datastores. Should I expect that CS will maintain one copy of each template
on all of the 4 datastores? If so, I'm wondering why CS is working this
way(Maybe I designed it wrong)?

Regarding the delay in VNC console, I tried to open a VNC session directly
from my PC to the VNC port that the ESX exposed for the VM and there was no
delay at  all. It seems that I experience this delay only when I'm using
the web console that works through the Console Proxy server. Can you
elaborate on how I can change the network throttle rate to 0? Is it with a
global parameter exposed through the UI? If so, which parameter is it?

Additional question, in the web UI I can see the ability to create only
anti-affinity-rules. Is there a way to create also affinity rules (I need
that a few VM's will be located on the same ESX host)? Is it exposed
through the API? If not, can you recommend on a plugin that can add this
option to CS?

Thanks again,
Lior.

On Tue, Jul 28, 2015 at 7:44 PM, ilya <ilya.mailing.li...@gmail.com> wrote:

> Lior,
>
> Since Rajani addressed #1 & #2,
> #3 is the issue you need to investigate on your side, as i cant reproduce
> this issue in many of my environments, some of them are very large and 2000
> miles away.
>
> But the flow is something along the lines of (if i recall correctly)
> 1) CloudStack contacts vCenter about VNC session request, vCenter responds
> with ESXi host and VNC port
> 2) CloudStack uses console proxy VM to establish a session directly to
> ESXi host VNC port and fronts it with web session to your browser over 443
> (or proxies the connection).
>
> Somewhere in this workflow - you have significant delay either on network
> or ESXi side. Also, consider change the default behavior of network
> throttle rate to 0, for some reason cloudstack default is 200 which limits
> the throughput on your network to 200mbps.
>
> Regards
> ilya
>
>
>
>
>
>
>
> On 7/28/15 12:09 AM, Rajani Karuturi wrote:
>
>> For 1: I am working on the following feature. This might be done in 4.7
>> timeframe. Please review and give your comments/suggestions
>>
>> https://cwiki.apache.org/confluence/display/CLOUDSTACK/LDAP%3A+Trust+AD+and+Auto+Import
>> For 2: Thats the expected behaviour
>>
>>
>> ~Rajani
>>
>> On Mon, Jul 27, 2015 at 11:34 PM, Lior Moneta <monetali...@gmail.com>
>> wrote:
>>
>>  Hi,
>>> I'm implementing CS in my environment and I encountered a few issues I
>>> couldn't find a clear answer to online.
>>>
>>> A little bit about my environment:
>>>
>>>     - I'm currently using one management server that holds both the
>>> MySQL DB
>>>     and the NFS for the secondary storage.
>>>     - I'm using vSphere 5.5 as my hypervisor.
>>>     - I've installed CloudStack 4.5.1.
>>>
>>> These are my questions:
>>>
>>>     1. *Microsoft LDAP integration* - I succeeded to integrate CS with
>>> my MS
>>>     AD but it seems that I have to manage all the users manually and I
>>> can't
>>>     find a way to manage it with AD groups. Is there a way to create a CS
>>>     account for a specific AD group? I would like to manage the users for
>>> the
>>>     accounts in the AD level and not manually in CS. I succeeded to
>>> create
>>> an
>>>     account that takes all the users from a specific group but it doesn't
>>>     update when I add a new user to the AD group.
>>>     2. *Duplicate for templates on each primary storage* - I have 2
>>> primary
>>>     storage (2 VMware vmfs datastores) and I noticed that If the template
>>>     doesn't yet exists on the datastore on which it deploys the VM, CS
>>> copy
>>> the
>>>     template from the secondary storage to that datastore although the
>>> template
>>>     already exist on the second datastore. Should it behave like that? Is
>>> there
>>>     a way to fix it?
>>>     3. *Console delay* - I'm experiencing about 1 second delay between my
>>>     actions and the response from the VM's console. This is an issue for
>>> me
>>>     since the VM's are going to be accessed only by the console window
>>> and
>>> not
>>>     by remote desktop or SSH. Wondering if this behavior is normal or is
>>> there
>>>     any configuration I need to implement in order to fix it. I'm
>>> experiencing
>>>     this delay both in Windows and Linux VM's with VMware Tools installed
>>> and
>>>     updated to the latest version.
>>>
>>> Thanks!
>>> Lior.
>>>
>>>
>

Reply via email to