For AttachVolume/DetachVolume API command, improve user error message in case of RuntimeException by throwing the exception instead of 'Unexpected Exception'.
(cherry picked from commit 4d7ede535df568c6aab4a228ac794ec11d433e1e)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Conflicts:
server/src/com/cloud/storage/VolumeApiServiceImpl.java
1. If a VM by the same name exists on a different cluster in the VMware DC, unregister the existing VM and continue with the VM start.
2. If VM start succeeds, delete VM files associated with the unregistered VM.
3. If VM start fails, re-register the unregistered VM.
(cherry picked from commit 974b0180dd67f19fea921092105161f849891ac5)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
If an exact match is being done while locating disk chain by name, don't trim snapshot postfix appended to the disk name.
(cherry picked from commit ddcae8a9306f5f7dd416b859e59fd10094c91551)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Consolidate VM disks once VM/volumes are migrated.
(cherry picked from commit cb211f18d14dcc9d988254a4b50b55ca0b080ed5)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Before registering a VM check if a different CS VM with same name exists in vCenter.
(cherry picked from commit 33179cce56b15f0632e38afa260cb829bb2a2273)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Look for a VM in vCenter based on both the vCenter name and CS internal name (required in case 'vm.instancename.flag' is enabled).
During Attach Volume and Volume Migration, for lookup and other operations use VM's name as obtained from vCenter instead of using the name set in the agent command.
(cherry picked from commit b8fdda0a34f5338db284ee616c5e04206c97b82e)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
During VM start while configuring its disk devices, obtain the matching disk for a volume in storage
using both the volume's path and volume's datastore information.
Change in vCenter 5.5 API from prior versions forced code change in CloudStack. Update property value of property "VirtualE1000.deviceInfo.summary" is accommodated now.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
[VMware] Synchronize the tasks of disk preparation and reconfiguration of a VM with the disk.
This is to avoid attaching a disk with a wrong device number on the controller key.
[VMware] While attaching a new disk to an instance, the unit number on the controller key should be the lowest unit number on the key that is not in use.
Instead of a number that is 1 digit higher the highest unit number that is currently in use.
vlan id format was like "vlan://<id>" instead of just "<id>". This causes numberformatexception while converting the vlan id to integer form from string form. this was fixed for standard vswitch in bug Cloudstack-5046. now fixed for other 2 cases of dvswitch as well as pvlan.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
In vCenter 5.5, once a volume is migrated the VMDKs are renamed to match the name of the VM.
If a volume has been renamed upon migration update its volumePath to that of the new disk filename.
Conflicts:
plugins/hypervisors/vmware/src/com/cloud/hypervisor/vmware/resource/VmwareResource.java
vmware-base/src/com/cloud/hypervisor/vmware/mo/VirtualMachineMO.java
We hit these excptions whenever a management server held session that was with the old 5.1 vCenter server
is used to make resource calls to the new 5.5 vCenter.
Validate a vCenter session context before it is being used to make a resource call.
And if the context is invalid then discard the context and retrieve a new one.
During the invalidation of an old context handle the context disconnect better
by catching the appropriate exception and returning a newly created context.
Conflicts:
plugins/hypervisors/vmware/src/com/cloud/hypervisor/vmware/resource/VmwareContextFactory.java
plugins/hypervisors/vmware/src/com/cloud/hypervisor/vmware/resource/VmwareResource.java
plugins/hypervisors/vmware/src/com/cloud/storage/resource/VmwareSecondaryStorageResourceHandler.java
vmware-base/src/com/cloud/hypervisor/vmware/util/VmwareContext.java
DetachISO is succeeding even though detach opeartion is failing as cdrom is locked by VM as it was mounted inside VM.
Detect if cdrom is locked or not. If locked fail detach operation and warn user to unmount before detaching the iso/cdrom device.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
Conflicts:
plugins/hypervisors/vmware/src/com/cloud/hypervisor/vmware/resource/VmwareResource.java
plugins/hypervisors/vmware/src/com/cloud/storage/resource/VmwareStorageProcessor.java
vmware-base/src/com/cloud/hypervisor/vmware/mo/VirtualMachineMO.java
During VM start, if VM already exists and CS finds a matching disk in storage for each of the existing disk then the VM is reconfigured with VM's existing disk information. This existing disk information for a VM is got from vCenter.
CS checks if a matching disk exists in storage by checking if there is a disk in storage whose name atleast partially matches with the name of the existing disk.
Post 4.1, Volume path has been changed to ROOT-<instanceId> from ROOT-<instanceId>-<volumeId>. Hence in case of an upgraded setup (pre 4.1 to 4.2) during search for an existing disk in storage even though ROOT-<instanceId>-<volumeId> has been deleted, CS does a postive match of the old existing disk because of the partial match with the new disk ROOT-<instanceId> that exists in storage. And so when VM is being powered on by vCenter it looks for ROOT-<instanceId>-<volumeId> and fails to power on.
Solution - While looking for a VM's matching disk in storage, instead of checking if a disk with a partial match exists check if a disk with the exact name exists.
