Mice, Yes, change 2 global configs you've mentioned + set system.vm.auto.reserve.capacity to false.
-Alena. On 7/11/12 8:20 PM, "Mice Xia" <mice_...@tcloudcomputing.com> wrote: >Hi, Alena, > >Im trying to follow your steps: > >* disable cluster >Succeed. > >* enable maintenance for the primary storage in the cluster >Maintenance on VMware cluster failed for the first two trys, with error >message like: >Unable to create a deployment for VM[ConsoleProxy|v-38-VM] > >WARN [cloud.consoleproxy.ConsoleProxyManagerImpl] (consoleproxy-1:) >Exception while trying to start console proxy >com.cloud.exception.InsufficientServerCapacityException: Unable to create >a deployment for VM[ConsoleProxy|v-47-VM]Scope=interface >com.cloud.dc.DataCenter; id=1 > >seems each time a new system VM was created, but still on VMware cluster, >which leads to failure >The maintenance succeed in the third try. > >* put hosts in cluster into maintenance mode >Succeed > >* destroy system vms >Destroying them does not stop them re-create > >* delete hosts and primary storage >Failed to delete primary storage, with message: there are still volumes >associated with this pool > >* delete the cluster > > >Putting hosts/storage into maintenance mode does not stop system VMs >re-create >From codes I can see management server get supported hypervisorTypes and >always fetch the first one, and the first one in my environment happens >to be vmware. > >I have changed expunge.interval = expunge.delay = 120 >Should I set consoleproxy.restart = false and update db to set >secondary.storage.vm=false ? > >Regards >Mice > >-----邮件原件----- >发件人: Alena Prokharchyk [mailto:alena.prokharc...@citrix.com] >发送时间: 2012年7月12日 10:03 >收件人: cloudstack-dev@incubator.apache.org >主题: Re: System VMs restarted on a disabled cluster > >On 7/11/12 6:29 PM, "Mice Xia" <mice_...@tcloudcomputing.com> wrote: > >>Hi, All >> >> >> >>I've set up an environment with two clusters (in the same pod), one >>Xenserver and the other is VMware, based on 3.0.x ASF branch. >> >>Now I'm trying to remove the VMware cluster begin with disabling it and >>destroying the system VMS running on it, but the systemVMs restarted >>immediately on VMware cluster, which blocks cluster removal. >> >> >> >>I wonder if this is the expected result by design, or should it be >>better that the system VMs get allocated on an enabled cluster? >> >> >> >> >> >>Regards >> >>Mice >> >> > > > >It's by design. Disabled cluster just can't be used for creating new / >starting existing user vms / routers; but it still can be used by system >resources (SSVM and Console proxy). > >To delete the cluster, you need to: > >* disable cluster >* enable maintenance for the primary storage in the cluster >* put hosts in cluster into maintenance mode > >* destroy system vms >* delete hosts and primary storage >* delete the cluster > >-Alena. > >
