Koushik Das created CLOUDSTACK-6402:
---------------------------------------
Summary: Fix StopCommand so that VMs are not removed accidentally
as part of vmsync
Key: CLOUDSTACK-6402
URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6402
Project: CloudStack
Issue Type: Bug
Security Level: Public (Anyone can view this level - this is the default.)
Components: Hypervisor Controller, Management Server
Affects Versions: 4.4.0
Reporter: Koushik Das
Assignee: Koushik Das
Fix For: 4.4.0
In order to prevent any side effects of out of band StopCommand issued by MS as
part of vmsync, some changes are required in the HV specific resource layer.
The StopCommand handler needs to be slightly changed as follows:
1. A new flag 'checkBeforeCleanup' will be introduced in the StopCommand
2. If the flag is not set (i.e. checkBeforeCleanup = false) then the existing
behaviour holds
3. If the flag is set (i.e. checkBeforeCleanup = true) then check status of VM
on the HV
a. If VM is running on HV then simply bailout without doing anything (return an
answer like StopAnswer(cmd, "VM is running on host, bailing out", false)
b. If not then do the same as step 2.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
