GabrielBrascher commented on issue #3600: Allow deploy Admin/System VMs in 
disabled zones/pods/clusters
URL: https://github.com/apache/cloudstack/pull/3600#issuecomment-585728562
 
 
   @andrijapanicsb thanks for raising those points. Sorry for the delay, it has 
been 4 months since I opened this PR and I needed to take a step back and look 
again to the use case, proposal, and implementation.
   
   > It seems that cloud admin can only start his own user VM on disabled 
resources - i.e. Admin can't start another user's user VM. But admin can start 
another user's VR.
   > Is this by design?
   
   You are right, the goal of this implementation is that admins can start 
their own VMs and System VMs; however, user VMs should not be deployed on 
disabled resources. It tackles a very specific case that we've seen in 
deprecated clusters/pods.
   
   >  I gave this a brief test, and it seems it's not working for system VMs 
(SSVM/CPVM)
   
   That is indeed related to the fact that there is only one host and the host 
is disabled. I did some tests and it works if disabling the cluster/pod/zone.
   
   > I disabled only the host (not zone/pod/cluster) - is this by design/OK - 
and should you just update the description of the global settings to also say 
"allow .. on disabled hosts" 
   
   The initial idea was indeed to allow hosts as well; however, there were some 
issues that would put the implementation on another level of complexity. It 
would also be a very specific corner case where all hosts are disabled, 
therefore it would make sense to disable the cluster/pod/zone instead of all 
hosts. That is why I described PR with disabled _zones/pods/clusters_.
   
   Additionally, from the implementation view: the code will mark storage pools 
as not suitable in case of local storage and disabled hosts. If there is no 
linkage of any suitable host (all hosts disabled) to any of the pools in the 
cluster, then the workflow considers the storage pool as "disabled" (not only 
hosts). Therefore, due to high complexity on solution for a very specific case, 
I decided to keep the implementation on zone/pod/cluster scope.
   
   > notice the missing "s" at the end of the 
"allow.system.vm.on.disabled.resource" vs " 
allow.admin.vm.on.disabled.resources"
   
   Going to fix that and also enhance documentation to let it clear that it 
works for disabled zone/pod/cluster :)

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

Reply via email to