[CLOUDSTACK-6121] Putting primary storage in maintenance mode does not shut down the VMs

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Not A Problem
• Affects Version/s: 4.3.0
• Component/s: KVM, Storage Controller, SystemVM
• Security Level: Public (Anyone can view this level - this is the default.)
• Labels: console kvm maintenance primary storage systemvm
• Environment: CentOS 6.5, KVM, NFS
• Assignee: Unassigned
• Reporter: Nux
• Created: 14/Feb/14 21:35
• Updated: 17/Feb/14 12:00
• Resolved: 17/Feb/14 12:00

Description

According to the docs, when putting the primary storage in maintenance mode all the VMs hosted on it shut be shutdown.

This does not happen in reality, the console proxy VM seems to be the only one that gets shut down but then it triggers a series of errors like:

2014-02-14 21:32:29,269 WARN [c.c.c.ConsoleProxyManagerImpl] (consoleproxy-1:ctx-5bc8688f) Exception while trying to start console proxy
com.cloud.exception.InsufficientServerCapacityException: Unable to create a deployment for VM[ConsoleProxy|v-1-VM]Scope=interface com.cloud.dc.DataCenter; id=1
(more here http://fpaste.org/77417/41354413/ )

It all gets stuck because of the system vm. `virsh list` is showing the other VMs running just fine.