[GitHub] [cloudstack] wido commented on pull request #4978: KVM High Availability regardless of storage

[GitBox <gi...@apache.org>]

wido commented on pull request #4978:
URL: https://github.com/apache/cloudstack/pull/4978#issuecomment-926556682


   > Ok, a couple of questions:
   > 
   > 1 - as Rohit asked, why can't we do the checks over SSH instead of running a separate service which might have security and other implications? Hypervisors are already expected to accept incoming SSH from the management server(s).
   
   No, we can't expect management servers always to SSH into the hypervisors. We for example have disabled this and the mgmt server can't SSH into the HV.
   
   This helper is there for the case that the CloudStack Agent crashes (for what reason) and the host is disconnected. Via this helper agent we have a second way of checking if the Host is still alive and has running VMs.
   
   If this helper does not respond either we can then fence off the host via OOB (IPMI or Redfish).
   
   Once the Host is Fenced we can safely start the VMs on a different host.
   
   > 2 - It's not 100% clear, people will still need to rely on the old NFS HA method, right? Your work merely adds additional checks. Any way we can get rid of NFS at this point?
   
   This additional method allows us to also provide HA when NOT using NFS. For example when using Ceph/RBD only.
   
   It is very, very, very much recommended to always use IPMI/Redfish with HA so that Hosts can be fenced properly.
   


-- 
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.

To unsubscribe, e-mail: commits-unsubscribe@cloudstack.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org