Sprint 39 review

ManageIQ
Sprint 39 Review - Sprint End April 19, 2016
April 20, 2016

Overview
● Sprint Statistics (Oleg Barenboim)
● User Interface (Dan Clarizio)
● Providers (Greg Blomquist)
● Platform (Gregg Tanzillo)
● REST API (Alberto Bellotti)
● SmartState (Rich Oliveri)
● Automate (Greg McCullough)
● Dockerizing ManageIQ (Barak Azulay)
● Performance (Dennis Metzger)
● Discussion

Sprint Statistics (O. Barenboim)
440 Pull Requests Merged (new record!)
* Note that some PRs
have more than one
label.

PR Breakdown by Feature Category
(O. Barenboim)
* Note that some PRs
have more than one
category.
Providers

Sprint Comparison (O. Barenboim)

All Repo Stats (O. Barenboim)
Repository PRs Merged
manageiq 440
ansible_tower_client 5
azure-armest 2
guides 5
manageiq_docs 1
manageiq-appliance 4
manageiq-appliance-build 4
manageiq-appliance-dev-setup 1
manageiq-ui-self_service 5
manageiq.org 5
miq_bot 2
ovirt 3

Darga is Coming
● Today, we are making the Darga branch
● Named after Klaus Darga (chess grandmaster)
● First Darga Build coming in next few days
● Next 2 sprints will be focused on stabilizing Darga
● Planning for Darga Release in June
● Name of “E” Release
○ Think of Candidate Names now
○ Vote during ManageIQ Summit (week of June 6)

User Interface (D. Clarizio)
● PRs merged (145+)
○ SSUI (5)
○ Bugs (64)
○ Enhancements (55)
○ Refactoring/Technical Debt (16)
● New functionality
○ New vertical navigation menus (Demo)
○ VM Reconfigure - add/remove disks support (Demo)
○ Orderable Orchestration Templates - create and copy (Demo)
○ Datastore Clusters for VMware vSphere 5+ - new Explorer
○ SlickGrid replaced with Patternfly TreeGrid
○ Patternfly style updates to the Dashboard and other areas
○ Template/Image compliance policy checking (previously only allowed for VMs/Instances)
○ New UI for replication configuration
○ UI updates to support multiple provider endpoints

UI - SlickGrid replaced with Patternfly TreeGrid

UI - Patternfly Style Updates
https://github.com/ManageIQ/manageiq/blob/master/app/assets/stylesheets/main.scss

UI - Demos
● New vertical navigation menus (E. Winchell)
● VM Reconfigure (H. Kataria)
● Orderable Orchestration Templates (H. Kataria)

Providers (G. Blomquist)
VMware
Azure (video)
OpenStack
Network Topology (video)
Hawkular Middleware
Containers
Ansible Tower
RHEV

VMware Clustered Datastores in Provisioning

Azure Subscription (@djberg96)

Azure Events (@bronaghs) [video]

OpenStack Ceilometer Events

Network Topology (@lsmola) [video]

Hawkular - Event Catcher
(H. Rupp)
Event Catcher
Currently for deployment error and success

Hawkular - Server reload
(H. Rupp)
Power operations to reload+stop MW-servers
Not reloading/stopping
the provider

Containers - Chargeback
(F. Simoncelli)

Containers - Chargeback
(F. Simoncelli)
● Chargeback totals per Project (Namespace)
○ CPU Cores usage
○ Memory usage
○ Network usage (when available)

Containers - OpenSCAP
(F. Simoncelli)
● Smart State Analysis has been extended with the
OpenSCAP support
○ Scanning Pod is instantiated in the container cluster
○ Regular SSA happens in ManageIQ using WebDAV to communicate to the
Scanning Pod
○ OpenSCAP scan runs in the Pod and ManageIQ collects the results
○ No interdependency between regular SSA and OpenSCAP
● At any time it’s possible to see the usual OpenSCAP
HTML results per Image
● Scans are scheduled at regular intervals and on a new
Image discovery
● Out-of-the-Box Policies for Image compliance
○ Additional ones can be created by the users
○ Results are pushed into OpenShift annotating Images

Containers - OpenSCAP
(F. Simoncelli)

Containers - Policies
(F. Simoncelli)
● Support introduced for OpenSCAP
○ Out-of-the-Box policies for Images SSA
● Can be extended further (Events and Policy entities) for
other use cases

Containers - Cloud Cross Linking
(F. Simoncelli)
● Already existing for Infrastructure Providers: RHEV,
VMWare, OpenStack
● Extended to Cloud Providers (GCE, AWS) using the
ProviderID reported by Kubernetes and OpenShift

Containers - Cloud Cross Linking
(F. Simoncelli)
● At the moment, requires a Cloud Provider
being configured in Kubernetes and
OpenShift (GCE, AWS)
● Ongoing effort in cAdvisor and Kubernetes
to automate the ProviderID
○ This will allow for other Cloud Providers to be
supported (Azure)
○ Targeted for OpenShift Origin 1.3

Containers - Dashboard Additions
(F. Simoncelli)
● Pods creation and deletion trends
● New images usage trend

Containers - Pods Network Metrics
(F. Simoncelli)
Add support for Pods Network Metrics
● Supported by Heapster in OpenShift Origin 1.2

Containers - More Enhancements
(F. Simoncelli)
● Upgrade to new Gems
○ Kubeclient 1.1.3 (was 0.8.1)
○ OpenShift Client 1.1.0 (was 0.3.0)
○ Image Inspector 1.0.1 (was 1.0.0)
● Display labels for Projects and Routes

Containers - Ongoing (F. Simoncelli)
● Continue OpenShift Deployment automation
● Use Multi-Endpoints for Hawkular Metrics
○ No additional steps to use OpenShift Metrics
○ No dedicated router on Master node
● Auto-Tagging based on Kubernetes Labels
○ It will be possible to map Regions and Zones to
specific Tags in ManageIQ
○ This can be extended to other Providers

Ansible Tower (G. McCullough)
● Model updates
○ Link JobTemplate instances to Inventory
○ Link Ansible hosts to VMs (when possible)
● New Automate namespace
○ /ManageIQ/ConfigurationManagement/AnsibleTower
● Demos
○ AnsibleTower as a Service Type Bill Wei
○ VM Custom Button Integration Madhu Kanoor
○ VM Provisioning Integration Madhu Kanoor

Generic Targeted Refresh Process
@blomquisg
RHEV Targeted Refresh
@pkliczewski

Multiple Endpoints (@juliancheal) [video]

Platform (G. Tanzillo)
Enhancements
● Replication
● Chargeback
● Appliance Console
● Rest API (A. Bellotti)

Platform - Replication
(G. Tanzillo)
Replication configuration screen
● Global and Remote regions
● Enable replication on remote regions
● Subscribe to remote regions on global

(G. Tanzillo)
Remote Region

(G. Tanzillo)
Global Region

Platform - Chargeback (G. Tanzillo)
Chargeback Rate Tiers

Platform - Appliance Console
(G. Tanzillo)
Appliance Console
● Added alias ‘ap’ as shortcut for appliance_console
○ alias ap='/usr/bin/appliance_console'

(A. Bellotti)
● Supports updating some external authentication settings
○ New menu entry and usage:

(A. Bellotti)
● External authentication options can also be updated using the Appliance
Console CLI:
# appliance_console_cli --extauth-opts=sso_enabled=true|false
# appliance_console_cli --extauth-opts=saml_enabled=true|false
# applicance_console_cli --extauth-opts=local_login_disabled=true|false

REST API (A. Bellotti)
● Versioning bumped to v2.2.0
● Added support for /api/settings
○ Provides access to subset of Settings
○ Driven by config/api_settings.yml
○ Supports GET /api/settings[/:section]
GET /api/settings/product
{
"product": {
"maindb": "ExtManagementSystem"
}
}

REST API (A. Bellotti)
● Now supporting CRUD on tenant quotas
● Create: POST /api/tenants/:id/quotas
{
“name” => “cpu_allocated”,
“value” => 1
}
● Queries: GET /api/tenants/:id/quotas
GET /api/tenants/:id/quotas/:quota_id
● Updates: POST /api/tenants/:id/quotas/:quota_id - action “edit”
PUT /api/tenants/:id/quotas/:quota_id
POST /api/tenants/:id/quotas - action “edit” to update multiple
● Deletes: POST /api/tenants/:id/quotas/:quota_id - action “delete”
DELETE /api/tenants/:id/quotas/:quota_id

SmartState (R. Oliveri)
SmartState Pull Requests:
● Total merged: 13
● Bugs: 9
○ SCVMM: 3
○ RHEV: 2
○ VMware: 1
○ Azure: 1
○ Common: 2
● Enhancements: 4

SmartState (R. Oliveri)
Features:
● LVM thin volume support.
○ Thin volumes ignored in the past.
○ VMs with root on thin volumes would fail.
○ VMs with thin volumes are now fully supported.
● SmartState Analysis for Azure.
○ Images
○ Instances
○ Optimization TBD

Automate (G. McCullough)
● Modeling changes
○ quota_source_type moved into instance
○ Added Auto-Approval/Email to VM Reconfigure
● VM Provisioning
○ Disable Auto-Placement logic for
create_provision_request and Rest API calls
■ Caller must provide all required placement fields
■ Default: placement_auto = true

Example:
"vm_fields" => {
"placement_auto" => "false",
"vm_name" => "test",
"Placement_availability_zone" => "2", # us-east-1b
"instance_type" => "2", # 'm1.small: M1 General Purpose Small'
"cloud_network" => "2",
"cloud_subnet" => "3",
"security_groups" => "64",
"floating_ip_address" => "4",
"guest_access_key_pair" => "65",
"monitoring" => "basic",
}

● Generic Object
○ New models
■ GenericObjectDefinition
● Defines the attributes for a GenericObject type.
● Relationships support will be added in future sprint
■ GenericObject
● Common attributes are saved in generic_objects table
○ name, uid and generic_object_definition
● All other attributes stored in custom_attributes
○ Next: Defining Object types and exposing to automate
● Automate - Git support
○ Modeling added
○ New “Git Repositories Owner” role
○ Branch/Tag lookup supported
○ Initial import logic
○ Next: rake task for import / UI support

Dockerizing ManageIQ (B. Azulay)
● What is this about
○ Be able to deliver ManageIQ docker appliance
○ To be able to install & run ManageIQ with a one liner command (on any OS that supports
docker, including Atomic)
● What can developers do with it
○ Deliver self built appliance for pre integration
○ Multiple environments on the same host (with separate builtin DBs)
○ Test your code in appliance mode
○ Test your code in a clean environment
● How should developers use it
○ Simply follow https://github.com/ManageIQ/manageiq/tree/master/docker-assets
● Things you should be aware of
○ Monolithic, Systemd, appliance mode, image size ~1.3 G
○ You can not run multiple manageiq containers at the same time on the same host (ports
collision)
○ First execution takes a bit longer (DB migrations run on first execution), simply wait till the
container listens to port 80 … and you are good (takes couple of minutes … depending
your HW)

Performance (D. Metzger)
Pull Requests:
● Total merged: 22
○ Bugs: 4
○ Enhancements: 18

Performance (D. Metzger)
● Support for sorting Virtual Columns in the database
○ Previously sorting on Virtual Columns was done in Ruby
● Metrics Rollups bug fix
○ Only load recent (needed) performance records not all records
● Service Tree improvement
○ Test at a customer site showed:
■ Reduction in page rendering time from 68 seconds to 10 seconds
■ Reduction in number of queries from 3,200 to 860
● RBAC
○ Ongoing effort to reduce the number SQL queries and quantity of data
being transferred

Discussion
Next Sprint Review - May 11

Sprint 39 review

More Related Content

What's hot (20)

Similar to Sprint 39 review (20)

More from ManageIQ (20)

Recently uploaded (20)

Sprint 39 review