There are some VM deployment failures happening when multiple VMs are deployed at a time, failures mainly due to NetworkModel code that iterates over all the vlans in the pod. This causes each deployVM thread to hold the global lock on Network longer and cause delays. This delay in turn causes more threads to choose same host and fail since capacity is not available on that host.
Following are some changes required to be done to reduce delays during VM deployments which in turn causes some vm deployment failures when multiple VMs are launched at a time.
In Planner, remove the clusters that do not contain a host with matching service offering tag. This will save some iterations over clusters that dont have matching tagged host
In NetworkModel, do not query the vlans for the pod within the loop. Also optimized the logic to query the ip/ipv6
In DeploymentPlanningManagerImpl, do not process the affinity group if the plan has hostId provided.
[4.10] CLOUDSTACK-7985: assignVM in Advanced zone with Security GroupsThis commit contains the following changes:
(1) implementation of assignVM in Advanced zone with Security Groups
(2) keep the default nic on shared network when assignVM
(3) allow migrate vm from/to project;
(4) UI change for selecting account/project/network
* pr/844:
CLOUDSTACK-7985: assignVM in Advanced zone with Security Groups
CLOUDSTACK-7985: keep the default nic on shared network when assignVM
CLOUDSTACK-7985: (1) allow migrate vm from/to project; (2) UI change for selecting account/project/network
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
- Bump spring-framework version to 4.x and Jetty to version that runs with JDK8
- Bump servet dependency version
- Migrate spring xmls to version 4, fixes schema locations that are 3.0
dependent in various xmls.
- Fix failing tests due to spring upgrade
(Thanks @marcaurele Marc-Aurèle Brothier for fixing them)
* Fix test DeploymentPlanningManagerImplTest
* Fix GloboDNS test
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Introduced a global configuration flag 'cluster.threshold.enabled'. By default the flag is true.
If the value is false, then a VM can be started in a cluster even if the cluster thresholds are
crossed. However, for a new VM deployment the cluster threshold will always be honoured.
- Improve disk chain usage while attaching, migrating disks
- Gets root disk controller based diskDeviceBusName from volume's chain info
- Refactor and move VirtualMachineDiskInfo to cloud-utils
- Allows mixing of scsi controller types
- Fixes a NPE case with map passed as null, for example in case of detach volume
command
- Use a osdefault translator that allow use of recent os types added (enums of
which) are not available in the sdk
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Earlier the deployment planner was not handling the case of virtual routers.(In Explicit Dedication)
It was only handling for all instance VMs/user VMs.
Added code for checking the case of Virtual Routers.
Added a source column to the user table.
Source now has only two values UNKNOWN,LDAP with UNKNOWN being the
default and is an enum is com.cloud.User.
When the source is UNKNOWN, the old method of authenticating against all
the available authenticators is used. If a source is available, only
that particular authenticator will be used.
added overloaded methods in AccountService to createUserAccount and
createUser with source specified.
(cherry picked from commit 5da733072e)
Design Document:
https://cwiki.apache.org/confluence/display/CLOUDSTACK/Allow+VM+listing+by+User+ID
- Adds column to VMInstance DAO
- Adds column in vm_instance table
- Adds column in the UserVMJoinVO
- Adds default admin user which has UID = 2
- Adds migration path that sets user_id to first user of the accountId that
owns the vm in vm_instance table
- Add arg on list VMs API to query by userId, add support in query layer
- Refactor VMInstanceVO and child classes to accept userId
- Add code to let service layer pass userId if loggedIn user belongs to same
account as the owner executing an API call or use first user from owner account
- In case of CPVM and SSVM use system user ID
- Fix unit tests and spring injections
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-4201: listServiceOfferings API needs to be able to take virtualmachineid of SystemVM and return service offerings available for the vm to change service offering
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
CLOUDSTACK-4762 : Enabling VGPU support for XenServer.
This feature is to enable the GPU-passthrough and vGPU functionality,
with the help of this feature, admins/users will be able to leverage
the GPU graphics unit power by deploying a virtul machine with GPU or
vGPU support or by changing the service offering of an existing VM
at any later point of time. There GPU/vGPU enabled VMs are able to run
graphical applications.
For now, this feature is only supported with XenServer hypervisor but
can be extended to add the support of other hypervisors.
Adding the missing file
During HA and maintenance call different planners (if the original planners are not able to find capacity) which skip some heurestics
Changes:
- Adding mocks in unit tests for new injected components
Conflicts:
server/test/org/apache/cloudstack/networkoffering/ChildTestConfiguration.java
There are three issues in resource_count table
(1) expunge a vm, the public_ip decreases and becomes -1 in basic zone.
(2) recover a vm, the volume increase.
(3) restore a vm, the volume decrease.
