ASF GitHub Bot commented on CLOUDSTACK-10310:

Slair1 commented on issue #2472: CLOUDSTACK-10310 Fix KVM reboot on storage 
URL: https://github.com/apache/cloudstack/pull/2472#issuecomment-372797350
   @wido My opinion is that the ability to start a VM on two hosts in the 
cluster is a bigger issue.  We run KVM and just enabled libvirt's lockd to stop 
that from being able to happen.  Maybe this PR could be changed to be an 
option, just reboot host or stop agent.

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

> KVM hosts reboot if there is a short transient storage error
> ------------------------------------------------------------
>                 Key: CLOUDSTACK-10310
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-10310
>             Project: CloudStack
>          Issue Type: Improvement
>      Security Level: Public(Anyone can view this level - this is the 
> default.) 
>          Components: KVM
>    Affects Versions: 4.9.0,
>            Reporter: Sean Lair
>            Priority: Major
> If the KVM heartbeat file can't be written to, the host is rebooted, and thus 
> taking down all VMs running on it.  The code does try 5x times before the 
> reboot, but the there is not a delay between the retires, so they are 5 
> simultaneous retries, which doesn't help.  Standard SAN storage HA operations 
> or quick network blip could cause this reboot to occur.
> Some discussions on the dev mailing list revealed that some people are just 
> commenting out the reboot line in their version of the CloudStack source.
> A better option (and a new PR is being issued) would be have it sleep between 
> tries so it isn't 5x almost simultaneous tries.  Plus, instead of rebooting, 
> the cloudstack-agent could just be stopped on the host instead.  This will 
> cause alerts to be issued and if the host is disconnected long-enough, 
> depending on the HA code in use, VM HA could handle the host failure.
> The built-in reboot of the host seemed drastic

This message was sent by Atlassian JIRA

Reply via email to