xs 6.1/6.2 introduce the new virtual platform, so there are two virtual platforms, windows PV driver version must match virtual platforms,
this patch tracks PV driver versions in vm details and template details.
Anthony
From functionality point of view following will be changes.
-> Threshold will not be applied for vms undergoing HA process (say during host going down event).
-> Threshold will not be applied for vms undergoing maintenance.
-> Threshold will apply for stop-start vm "after" the skip.counting.hours
(time till which we reserve the capacity for the stopped vm).
Reason being stop-start is a user initiated action and we are giving a time
window (skip.counting.hours) till which its guaranteed to start in the same
cluster.
After that time, threshold check applies bcz we still want to reserve the
capacity for urgent situations like HA, putting host into maintenance rather than get it consumed in such user
initiated actions.
Signed off by : nitin mehta<nitin.mehta@citrix.com>
Applying the short term fix of force cleaning up if the answer recieved from startcommand is not valid
Signed off by : nitin mehta<nitin.mehta@citrix.com>
Updating the fix to cover one more scenario when user directly calls API migrateVirtualMachineWithVolume.
If currentPool is accessible to destination host, skip calling allocators and move on to next volume to process.
This means if user calls migrateVirtualMachineWithVolume API where all volumes of VM are accessible on specified target host,
then API fails as there is no storage migration involved. Instead user should call migrateVirtualMachine API.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
This is due to a VM on zone wide primary storage not requiring storage migration while migrating across clster.
Detecting the storage pool type before allowing normal migration (without storage live migration) of VM across cluster.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
Changes:
- Passing the avoid set generated by the first pass of deployment to the second try.
- The second try is done, when the first pass that uses a reserved plan fails to deploy on the reserved host, to search over the entire zone again
The problem was because in cloudstack when a vm is stopped it gets destroyed on the host. For a
windows vm the timeoffset (which can be set by changing the timezone from within the vm) is stored
in the platform:timeoffset attribute of vm record. The information is lost when the vm is destroted.
Made change to read and persist the platform:timeoffset vm attribute when an instance is stopped.
The value is persisted in the user_vm_details table. When the vm is started again the attribute is
set for the vm instance that gets created.
If the scaling up fails on the host the vm is running on try to migrate it to other hosts in the cluster and try scaling.
CLOUDSTACK-3349
For deciding the host in the cluster try the new deployment manager now
Signed off by : nitin mehta<nitin.mehta@citrix.com>
CLOUDSTACK-3042 - handle Scaling up of vm memory/CPU based on the presence of XS tools in the template
This also takes care of updation of VM after XS tools are installed in the vm and set memory values accordingly to support dynamic scaling after stop start of VM
Signed-off-by: Abhinandan Prateek <aprateek@apache.org>
for number of commands participating in Vm deployment process, as parallel deployment is supported on the hypervisor side.
The behavior is controlled by global config varirables:
"execute.in.sequence.hypervisor.commands" (false by default) sets/resets the synchronization for commands:
=========================
StartCommand
StopCommand
CreateCommand
CopyVolumeCommand
"execute.in.sequence.network.element.commands" (false by default) sets/resets the synchronization for commands:
==========================
DhcpEntryCommand
SavePasswordCommand
UserDataCommand
VmDataCommand
As a part of the fix, increased the global lock timeout to 30 mins in several VR scripts:
===========================
edithosts.sh
savepassword.sh
userdata.sh
to support situations when multiple concurrent calls to the script are being made.
During Scale up of VM, memory/cpu calculations should consider the memory/cpu overprovisioning factors which are set per cluster.
CLOUDSTACK-2939: CPU limit is not getting set for vm after scaleup to a service offering which have cpu cap enabled
Signed-off-by: Abhinandan Prateek <aprateek@apache.org>
The location of the virtual machine is provided by DeployDestination, which will
be passed in NetworkGuru#reserve and NetworkElement#prepare.
During the virtual machine migration, it actually changes DeployDestination and
it looks like that it will tell that event to network components as it has
NetworkManager#prepareNicForMigration. The problem is that althogh the interface
has that method, NetworkManagerImpl does not tell the DeployDestination changes
to network components.
So IMHO, we need to add calls of NetworkGuru#reserve and NetworkElement#prepare
in NetworkManagerImpl#prepareNicForMigration . And then, we also need to add
calls NetworkGuru#release and NetworkElement#release after the migration,
otherwise the network resources that plugin reserved will be kept even when the
vm leaves off.
(Sheng Yang: rebase code, add license header)
Signed-off-by: Sheng Yang <sheng.yang@citrix.com>
1. Keeping the description consistent - Memory not RAM when referring to
overcommit
2. getters And setters grouped, provided right casing.
3. Removed wildcard imports
Signed-off-by: Prasanna Santhanam <tsp@apache.org>
Changes:
- There is no good mechanism currently to figure out if the deployment failed due to affinity groups only
- We can just hint the user that the deployment might have failed due to the affinity groups and ask to review the input
Wei Zhou
Anthony Xu
Nitin Mehta
Edison Su
Kelven Yang
Sateesh Chodapuneedi
Alena Prokharchyk
Bharat Kumar
Kishan Kavala
Harikrishna Patnala
Prachi Damle
Hugo Trippaers
Devdeep Singh
Edison Su
Min Chen
Mice Xia
Wido den Hollander
Hiroaki Kawai
Prasanna Santhanam
Alex Huang
Sanjay Tripathi
Marcus Sorensen