>> start VM which goes to hypervisor hosts
>
> Not directly to the hypervisors right? They go through vCenter if I
> understand correctly.

Right, the commands go through vCenter, similar to commands to XS hosts
going through the pool master.  But the command sequences will be
host-specific, which you can see in the the logs, the commands can get
stuck waiting for previous ones sent to the same host.

Does your 4.1 environment have the following Global Settings, and if so,
what are the values?

execute.in.sequence.hypervisor.commands
execute.in.sequence.network.element.commands

If they are set to false, try setting them to true to see if the
behavior improves.

Best regards,
Kirk


On 09/10/2014 03:27 AM, Ian Duffy wrote:
> Hi Kirk,
> 
>> Did you figure this out?
> No, we didn't look into it much more. Its not causing issues just yet and
> is just a curiosity thing.
> 
>> start VM which goes to hypervisor hosts
> 
> Not directly to the hypervisors right? They go through vCenter if I
> understand correctly.
> 
>> How many VMware hosts do you have?
> 
> We have 2 ESXi hosts.
> 
> Thanks,
> 
> Ian
> 
> 
> On 10 September 2014 11:19, Kirk Kosinski <[email protected]> wrote:
> 
>> Hi, Ian.  Did you figure this out?  Template deployment in VMware is
>> done through the SSVM instead of hosts so the behavior may be different
>> than other commands, like start VM which goes to hypervisor hosts.  How
>> many VMware hosts do you have?  If you have few or just one VMware host,
>> the commands are probably queuing up (check the logs for "Waiting for
>> Seq").
>>
>> Best regards,
>> Kirk
>>
>>
>> On 09/05/2014 03:28 AM, Ian Duffy wrote:
>>> Hi All,
>>>
>>> I'm using Cloudstack 4.1.1 against vCenter 5.
>>>
>>> If I create 5 instances on cloudstack. I see the 5 disks get created in
>>> parallel but the power up and power down events do not.
>>>
>>> Is this expected functionality? Or should power up and power down events
>>> occur at the same time? If so is there some configuration around this?
>>>
>>> If I go into vCenter I can manually power up and power down machines in
>>> parallel without issue.
>>>
>>> Thanks,
>>>
>>> Ian
>>>
>>
> 

Reply via email to