Sprint 125
1 like120 views
The sprint review covered improvements across several areas including the user interface, providers, platform, API, and quality engineering. Key enhancements included adding code owners for provider repositories, allowing tagging of resources, and improving metrics capture. Bugs addressed included image refresh issues and workers not showing correctly. Testing coverage was expanded through new tests and framework improvements.
Sprint 125
- 1. SPRINT 125 REVIEW - Sprint Ending November 25, 2019 November 27, 2019 Meeting Start: 10:30 am EST
- 2. 2 OVERVIEW ● UI (Harpreet Kataria) ● Providers (Adam Grare) ● Platform (Joe VLcek) ● API (Martin Povolny) ● QE (Mike Shriver) ● Summary
- 3. 3 USER INTERFACE (Harpreet Kataria) PRs (UI, SUI, UI/REACT Components, V2V): 55 Bugs: 22, Enhancements: 3, Technical Debt/cleanup: 10/10, Test: 3, others: 7 Enhancements ● Add page title for custom menu items (#6387) ● Allow tagging or managing policies for Replicators, Pods, Nodes, Images (#6379) ● Add Refresh button to individual Dashboard Widget (#6334) Bugs ● DialogUser - update validator_type during field refresh (#428) ● v2v: ScheduledTimeInfoItem: fix PropType for migrationScheduled (#1067) ● Fix cancel/reset buttons' reaction to C&U collection checkboxes (#6432) ● Compare Treegrid fix (#6417) ● Call `miqCheckForChanges` before changing tabs. (#6431) ● Add fallback for clicking on a section w/o permissions. (#6414) ● Display user info after clicking on Lookup while provisioning VM (#6391) Classic UI Bug fixes
- 4. 4 PROVIDERS (Adam Grare) ● Add code owners for all provider repositories ○ app/models/manageiq/providers* and every provider plugin now has at least two people as owners to review PRs ○ If you want to help out please add yourself! ● Remove VMwareWebService/MiqVimVm from Core (#19551) ○ A number of “special” snapshot name constants were being used in core from the VimBroker ○ The EVM snapshot name should be moved to core anyway ○ The others (VCB and Consolidate Helper) should only be used in the VMware repo ○ In addition, the logic for if a snapshot should be removed was moved out of core and into the VMware repo ● Subclass ResourcePool by provider (#19529) ○ Create a sub-class for Ovirt, SCVMM, and VMware
- 5. 5 PROVIDERS (Adam Grare) ● Remove ems_ref_obj from a number of models (#19430) ○ We had a YAML serialized column on e.g. VMs, Hosts, EmsClusters, etc… ○ This was to store the entire ManagedObjectReference for VMware (basically VIM class + ems_ref) ○ This required us to have the VMwareWebService gem in core to YAML.dump this column ○ Replaced with an ems_ref_type to store the VIM class, VMware provider builds the MOR ● Add _queue methods for all VM operations (#19544) ○ Add a common method to queue VM operations removing some duplications ○ Allow a queue_name to be set for all VM operations (enable some future changes) ○ Don’t call raw_ from check_policy_prevent, allowing the operation to be run on a different role from “automate” (enabling the same future changes)
- 6. 6 PROVIDERS (Adam Grare) ● Move perf_capture to ...Providers::BaseManager::MetricsCapture class (#19541) ○ Any logic in the base Metrics::* classes cannot be overridden by the different providers ○ MetricsCapture is sub-classed by each provider which allows us to modify the behavior of metrics capture to enable performance enhancements ● Extract all of the raw_ VM operations to providers (#19452) ○ Some providers have the guts of the VM operations methods defined on their manager ○ The base VmOrTemplate::Operations raw_* methods all defined run_command_via_parent ○ Most of the provider overrode these so having them defined in core was confusing ○ Now the core raw_ methods raise Unimplemented exceptions and are defined in the providers
- 7. 7 PROVIDERS (Adam Grare) OpenStack ● Fix image targeted refresh (#521) ○ Oslo events had the image id in a different attribute in the event data ● Fix nested stack targeted refresh (#534) ○ Events for nested stacks weren’t including their parent stacks so inherited resources weren’t being connected properly
- 8. 8 PLATFORM (Joe VLcek) 52 PRs merged Enhancements: ● Add initial codeowners file container-embedded-ansible/7 container-kafka/1 container-memcached/3 container-postgresql/16 container-ruby/15 manageiq-appliance/263 manageiq-pods/349 manageiq-postgres_ha_admin/18 pg-dsn_parser/6 pg-logical_replication/3 pg-pglogical/25 manageiq/18594 ● metrics_capture supports enhancements manageiq/19511 manageiq/19506 manageiq/19522 ● Adding support for configuring Appliance SAML Authentication via the CLI manageiq-appliance_console/102 manageiq/19525 manageiq-appliance/264 ● Kubernetes-ify the container orchestrator manageiq/19500 manageiq-pods/348 ● Add MiqWorkerType model manageiq/19536 manageiq-schema/437 ● Replace the server capabilities serialized column with has_vix_disk_lib manageiq-schema/439 ● Add Class#leaf_subclasses method more_core_extensions/71 ● Rename table chargeback_rate_detail_currencies to currencies manageiq-schema/421 ● Use table currencies if exists in ChargebackRateDetailCurrency manageiq/19350 ● Add currency model and factory manageiq/19542 ● Update linux_admin to 1.2.4 manageiq/19493 ● Ongoing test coverage improvements ○ fix virtual spelling mistake activerecord-virtual_attributes/56 ○ Add tests for match_query_limit_of activerecord-virtual_attributes/55 ○ fix sporadic test with state_machine_spec manageiq/19530
- 9. 9 PLATFORM (Joe VLcek) Bugs: ● Remove ovirt V3 api calls manageiq/19508 ● Use the workers proc rather than overriding the method manageiq/19494 ● Allow dashboards for different groups to have the same name manageiq/19491 manageiq-schema/432 ● Display available disk in utilization manageiq/19502 ● Don't convert user-type email to upn manageiq/19515 Technical Debt/Refactoring: ● Use parent as the default implementation of .ems_class manageiq/19516 manageiq/19513 ● Remove the Runner#message_restarted method manageiq/19545 ● Remove repmgr10 from the docker image manageiq/19505 ● Remove server capabilities serialized column manageiq/19552 ● Address workers not present in worker list manageiq/19524 ● metrics simplify calc_targets_by_rollup manageiq/19518 ● introduce capture interval_priority manageiq/19519 ● Cu capture object zone manageiq/19541 ● Update MiqExpression#registry_data and add coverage manageiq/19497 ● Hawkular removal manageiq/19378 ● Change NoMethodError to NotImplementedError in MiqRequestWorkflow manageiq/19554
- 10. 10 PLATFORM (Joe VLcek) Continuous Integration / Testing ● Clean up extra space used by gems and bundler manageiq-appliance-build/368 ● Load graphql only in puma process bundler groups (api/ui) manageiq/19517 ● Use secret_key_base to avoid rails 5.2 deprecation manageiq/19514
- 11. 11 API (Martin Povolny) ● Improvements to specs by (Keenan Brock) ● Protect user’s changes to PXE Images when editing PXE Servers #707 (Miha Pleško)
- 12. 12 QE (Mike Shriver) ● https://github.com/ManageIQ/integration_tests/releases/tag/17.63.0 ● New 2nd level reviewer, Nikhil Dhandre! (http://github.com/digitronik) ● fix-test 20 ● test-automation 15 ● enhancement 10 ● rc-regression-fix 8 ● fix-framework 7 ● new-test-or-feature 5 ● test-cleanup 5 ● fix-locator-or-text 3 ● other 3 ● manual 3 ● customer-case 2 ● tech-debt 2
- 13. 13 SUMMARY Sprint 126 Review - December 11, 2019 Meeting Start: 10:30 am EST
- 14. THANK YOU manageiq.org github.com/ManageIQ twitter.com/ManageIQ gitter.im/ManageIQ/manageiq facebook.com/manageiq talk.manageiq.org plus.google.com/+ManageiqOrg youtube.com/user/ManageIQVideo linkedin.com/company/manageiq