VM dynamic scaling option granularity by harikrishna-patnala · Pull Request #4643 · apache/cloudstack

harikrishna-patnala · 2021-02-03T05:51:27Z

Description

This PR introduces new granularity levels to configure VM dynamic scalability. Previously VM is configured to be dynamically scalable based on the template and global setting. Now we bringing this option to configure at service offering and VM level also.

VM can dynamically scale only when all flags are ON at VM level, template, service offering and global setting. If any of the flags is set to false then VM cannot be scalable. This result will be persisted in DB for each VM and will be honoured for that VM till it is updated.

We are introducing 'dynamicscalingallowed' parameter with permitted values of true or false for deployVM API and createServiceOffering API.

Following are the API parameter changes:
createServiceOffering API:
dynamicscalingenabled: an optional parameter of type Boolean with default value “true”.
deployVirtualMachine API:
dynamicscalingenabled: an optional parameter of type Boolean with default value “true”.

Following are the UI changes:
Service offering creation has ON/OFF switch for dynamic scaling enabled with default value true.

Deploy VM wizard has ON/OFF switch for dynamic scaling enabled with default value true in Advanced Mode section.

Added a documentation PR to support these changes
apache/cloudstack-documentation#186

Types of changes

Breaking change (fix or feature that would cause existing functionality to change)
New feature (non-breaking change which adds functionality)
Bug fix (non-breaking change which fixes an issue)
Enhancement (improves an existing feature and functionality)
Cleanup (Code refactoring and cleanup, that may add test cases)

Feature/Enhancement Scale or Bug Severity

Feature/Enhancement Scale

Major
Minor

Bug Severity

Screenshots (if appropriate):

How Has This Been Tested?

Following are the test cases that I've covered as part of my testing on VMware environment.

VM Creation:

Created VM with Small service offering with all Flags ON
- VM1 (scalable)
Created VM with Small service offering which is not dynamic, with Dynamic Scaling Enabled true on VM
- VM2 (not scalable)
Created VM with Small service offering, with Dynamic Scaling Enabled false on VM
- VM3 (not scalable)
Created VM with Small service offering which is not dynamic, with Dynamic Scaling Enabled false on VM
- VM4 (not scalable)
Created VM with Small service offering with all dynamic flags ON except template
- VM5 (not scalable)

VM Update:

Stop VM1 => set Dynamic Scaling Enabled to false => Update VM
- VM is updated to Dynamic Scaling Enabled value false
Stop VM2 or VM3 or VM4 or VM5 => set Dynamic Scaling Enabled to true => Update VM
- VM update operation fails because VM cannot be scaled with any of the template/service offering/global setting dynamic scaling enabled flag with false

VM dynamic Scaling:

While trying to scale VM from UI, it will only list service offering that are suitable to scale and also offering with "Dynamic Scaling Enabled" value same as old service offering.

Create Big Service Offering with Dynamic Scaling Enabled true
- SO1
Create Big Service Offering with Dynamic Scaling Enabled false
- SO2
Scale VM1 (scalable) to Service Offering SO1
- Successful
Scale VM1 (scalable) to Service Offering SO2
- Operation failed because, VM1 Service Offering is marked Dynamic Scaling Enabled but not on SO2

harikrishna-patnala · 2021-02-03T05:56:30Z

@blueorangutan package

api/src/main/java/org/apache/cloudstack/api/ApiConstants.java

api/src/main/java/org/apache/cloudstack/api/response/ServiceOfferingResponse.java

yadvr · 2021-02-03T10:48:40Z

@blueorangutan ui

harikrishna-patnala · 2021-02-04T10:03:32Z

@blueorangutan package

harikrishna-patnala · 2021-02-04T10:44:20Z

@blueorangutan test matrix

harikrishna-patnala · 2021-02-05T06:56:18Z

These failed test cases are commonly seen on all PRs, not related to this PR changes.

test_01_scale_vm on XenServer is failing with License error

harikrishna-patnala · 2021-02-05T06:58:21Z

@blueorangutan ui

utchoang

UI Code LGTM!

...src/main/java/org/apache/cloudstack/api/command/admin/offering/CreateServiceOfferingCmd.java

Pearl1594 · 2021-02-05T08:52:47Z

engine/schema/src/main/java/com/cloud/service/ServiceOfferingVO.java


+    @Column(name = "dynamic_scaling_enabled")
+    private boolean dynamicScalingEnabled;
+


@harikrishna-patnala what's the difference between "dynamic_scaling_enabled" vs "is_dynamic" field?

is_dynamic is related to custom compute offerings and thats a transient parameter.

Pearl1594 · 2021-02-05T08:56:22Z

server/src/main/java/com/cloud/network/as/AutoScaleManagerImpl.java

                    getCurrentTimeStampString(),
                    "autoScaleVm-" + asGroup.getId() + "-" + getCurrentTimeStampString(), null, null, null, HypervisorType.XenServer, HTTPMethod.GET, null, null, null,
-                    null, true, null, null, null, null, null, null, null);
+                    null, true, null, null, null, null, null, null, null, true);


@harikrishna-patnala is there a reason that we've set it to true as opposed to find the value from the service offering?

this parameter value in deploy VM flow is the one that read from deploy VM API, where as in other cases there is no user/admin option to specify whether the VM can be dynamically scalable or not, so it is defaulted to true. All these method calls will reach checkIfDynamicScalingCanBeEnabled() where other dynamic scaling flags from template, offering and global setting are considered.

Pearl1594 · 2021-02-05T08:56:31Z

server/src/main/java/com/cloud/network/as/AutoScaleManagerImpl.java

                        owner, "autoScaleVm-" + asGroup.getId() + "-" + getCurrentTimeStampString(),
                        "autoScaleVm-" + asGroup.getId() + "-" + getCurrentTimeStampString(), null, null, null, HypervisorType.XenServer, HTTPMethod.GET, null, null,
-                        null, null, true, null, null, null, null, null, null, null);
+                        null, null, true, null, null, null, null, null, null, null, true);


same as above

Pearl1594 · 2021-02-05T08:56:39Z

server/src/main/java/com/cloud/network/as/AutoScaleManagerImpl.java

                    vm = _userVmService.createAdvancedVirtualMachine(zone, serviceOffering, template, null, owner, "autoScaleVm-" + asGroup.getId() + "-" +
                        getCurrentTimeStampString(), "autoScaleVm-" + asGroup.getId() + "-" + getCurrentTimeStampString(),
-                        null, null, null, HypervisorType.XenServer, HTTPMethod.GET, null, null, null, addrs, true, null, null, null, null, null, null, null);
+                        null, null, null, HypervisorType.XenServer, HTTPMethod.GET, null, null, null, addrs, true, null, null, null, null, null, null, null, true);


same as above

server/src/main/java/com/cloud/vm/UserVmManagerImpl.java

nvazquez

Nice feature, I left some comments

nvazquez · 2021-02-06T13:55:43Z

...src/main/java/org/apache/cloudstack/api/command/admin/offering/CreateServiceOfferingCmd.java

    }

+    public Boolean getDynamicScalingEnabled() {
+        return isDynamicScalingEnabled == null ? Boolean.TRUE : isDynamicScalingEnabled;


As the null check is in place, what about returning boolean instead of Boolean on this function? Also, this will mean that all the service offerings registered before 4.16 will now become dynamically scallable as well?

fixed all Boolean to boolean, thanks for that.
Yes all old service offerings prior to upgrade will be marked as dynamically scalable. This is made not to break existing VMs because service offering was not part of dynamic scaling decision before. After upgrade we need to bypass that check for old VMs, which can be achieved by marking them true.

Copy, thanks

@harikrishna-patnala and @nvazquez this method is Boolean was it meant to become boolean?

api/src/main/java/org/apache/cloudstack/api/command/user/vm/DeployVMCmd.java

engine/schema/src/main/java/org/apache/cloudstack/engine/cloud/entity/api/db/VMEntityVO.java

server/src/main/java/com/cloud/vm/UserVmManagerImpl.java

harikrishna-patnala · 2021-02-08T11:08:05Z

@sureshanaparti @Pearl1594 @nvazquez I've addressed your comments and updated the code as per the suggestions. Please review.

harikrishna-patnala · 2021-02-08T11:32:57Z

@blueorangutan package

blueorangutan · 2021-02-08T11:33:57Z

@harikrishna-patnala a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress.

...src/main/java/org/apache/cloudstack/api/command/admin/offering/CreateServiceOfferingCmd.java

api/src/main/java/org/apache/cloudstack/api/command/user/vm/DeployVMCmd.java

engine/schema/src/main/java/com/cloud/vm/VMInstanceVO.java

server/src/main/java/com/cloud/vm/UserVmManagerImpl.java

…tiple checks Added the case for restore VM where VM can be restored to new template. In that case VM's dynamic scalability needs to be updated based on new template

…ceOfferingVO

…d as dynamically scalable.

…ce offering and template are not dynamically scalable

… be dynamically scalable

…s not enabled on template, service offering and global setting.

harikrishna-patnala · 2021-05-28T10:26:13Z

@blueorangutan package

blueorangutan · 2021-05-28T10:27:02Z

@harikrishna-patnala a Jenkins job has been kicked to build packages. I'll keep you posted as I make progress.

blueorangutan · 2021-05-28T10:56:14Z

Packaging result: ✔️ centos7 ✔️ centos8 ✔️ debian. SL-JID 110

harikrishna-patnala · 2021-05-28T11:10:29Z

@blueorangutan test matrix

DaanHoogland · 2021-05-29T11:28:45Z

@blueorangutan test matrix

blueorangutan · 2021-05-29T11:29:03Z

@DaanHoogland a Trillian-Jenkins matrix job (centos7 mgmt + xs71, centos7 mgmt + vmware65, centos7 mgmt + kvmcentos7) has been kicked to run smoke tests

blueorangutan · 2021-05-29T21:29:51Z

Trillian test result (tid-788)
Environment: kvm-centos7 (x2), Advanced Networking with Mgmt server 7
Total time taken: 33951 seconds
Marvin logs: https://github.com/blueorangutan/acs-prs/releases/download/trillian/pr4643-t788-kvm-centos7.zip
Intermittent failure detected: /marvin/tests/smoke/test_internal_lb.py
Smoke tests completed. 88 look OK, 0 have error(s)
Only failed tests results shown below:

Test	Result	Time (s)	Test File

alexandremattioli · 2021-06-01T11:39:05Z

LGTM

yadvr · 2021-06-01T11:40:22Z

@blueorangutan test centos7 vmware-67u3

blueorangutan · 2021-06-01T11:41:02Z

@rhtyd a Trillian-Jenkins test job (centos7 mgmt + vmware-67u3) has been kicked to run smoke tests

yadvr · 2021-06-01T11:57:59Z

@blueorangutan test centos7 xenserver-71

blueorangutan · 2021-06-01T11:59:02Z

@rhtyd a Trillian-Jenkins test job (centos7 mgmt + xenserver-71) has been kicked to run smoke tests

blueorangutan · 2021-06-01T12:22:35Z

Trillian Build Failed (tid-815)

blueorangutan · 2021-06-01T12:45:31Z

Trillian Build Failed (tid-814)

blueorangutan · 2021-06-04T19:31:35Z

Trillian test result (tid-858)
Environment: xenserver-72 (x2), Advanced Networking with Mgmt server 7
Total time taken: 37475 seconds
Marvin logs: https://github.com/blueorangutan/acs-prs/releases/download/trillian/pr4643-t858-xenserver-72.zip
Intermittent failure detected: /marvin/tests/smoke/test_list_ids_parameter.py
Intermittent failure detected: /marvin/tests/smoke/test_volumes.py
Smoke tests completed. 88 look OK, 0 have error(s)
Only failed tests results shown below:

Test	Result	Time (s)	Test File

blueorangutan · 2021-06-04T19:41:48Z

Trillian test result (tid-859)
Environment: vmware-67u3 (x2), Advanced Networking with Mgmt server 7
Total time taken: 37368 seconds
Marvin logs: https://github.com/blueorangutan/acs-prs/releases/download/trillian/pr4643-t859-vmware-67u3.zip
Intermittent failure detected: /marvin/tests/smoke/test_list_ids_parameter.py
Smoke tests completed. 88 look OK, 0 have error(s)
Only failed tests results shown below:

Test	Result	Time (s)	Test File

harikrishna-patnala marked this pull request as draft February 3, 2021 05:55

harikrishna-patnala added this to the 4.16.0.0 milestone Feb 3, 2021