[
https://issues.apache.org/jira/browse/CLOUDSTACK-5578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14215390#comment-14215390
]
edison su commented on CLOUDSTACK-5578:
---
It's due to the behavior on kvm host:
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Virtualization_Deployment_and_Administration_Guide/sect-Managing_guest_virtual_machines_with_virsh-Shutting_down_rebooting_and_force_shutdown_of_a_guest_virtual_machine.html
Normally, KVM host will try to suspend vms on host during reboot, which may
stuck when primary storage is unavailable.
As you said, we could try "reboot -f", or change the KVM host behavior, to
forcefully shutdown vms.
> KVM - Network down - When the host looses network connectivity , reboot stuck
> while unmounting primary
> --
>
> Key: CLOUDSTACK-5578
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-5578
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: Management Server
>Affects Versions: 4.2.0
> Environment: Build from 4.3
>Reporter: Sangeetha Hariharan
>Assignee: Kishan Kavala
>Priority: Critical
> Fix For: 4.5.0
>
> Attachments: DisconnectedHost.png, kvm-hostdisconnect.rar,
> nfsUmount.jpg
>
>
> KVM - Network down - When the host looses network connectivity , it is not
> able to fence itself.
> Steps to reproduce the problem:
> Set up - Advanced zone with 2 Rhel 6.3 hosts in cluster.
> Deploy ~10 Vms.
> Simulate network disconnect on the host ( ifdown em1)
> Host gets marked as "Down" and all the Vms gets HA-ed to the other host.
> On the KVM host which lost connectivity , attempt to shutdown itself fails.
> It was not able to umount the primary store.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)