wido commented on pull request #4978:
URL: https://github.com/apache/cloudstack/pull/4978#issuecomment-836766195


   > HI, I see this PR is very good.
   > 
   > I have a question, if the manager network link is interrupted (such as a 
switch failure), but the storage network of kvm is still running normally.
   > 
   > Will this cause the VM to trigger HA?
   > 
   > I am worried that the image is "double written", resulting in damage to 
the image
   
   You should always have the Out of Band management of the hypervisors 
configured:
   
   - IPMI 
   - Redfish
   
   When the mgmt server wants to perform HA it will Fence off  the Host by 
performing a power cycle of that host via OOB.
   
   This will make sure no VMs are running and they can safely be started on a 
different host.
   
   In addition we also have locking mechanisms:
   
   - file locks on QCOW2 when using NFS (modern Qemu required)
   - exclusive locking with RBD/Ceph on images


-- 
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:
[email protected]


Reply via email to