duplicate record was detected and destroyed, inventory_collection: 'InventoryCollection:<CustomAttribute>

ManageIQ / manageiq-providers-vmware

ManageIQ plugin for the VMware vSphere and vCloud providers.

Apache License 2.0

22 stars 70 forks source link

duplicate record was detected and destroyed, inventory_collection: 'InventoryCollection:<CustomAttribute> #823

Closed szajac01 closed 2 years ago

szajac01 commented 2 years ago

Hi After migrating from kasparov-1 to kasparov-2, we have a lot of message like this in evm.log

<[----] W, [2022-09-09T11:33:26.087553 #2774:2ae19134b1b8] WARN -- : A duplicate record was detected and destroyed, inventory_collection: 'InventoryCollection:, strategy: local_db_find_missing_references', record: '#', duplicate_index: 'Some Index Names'

A series of message occurs every 3 or 5 minutes, we want to understand what is this problem before going on production

thanks

Fryguy commented 2 years ago

@agrare I believe this is fixed already, but please take a look.

agrare commented 2 years ago

This was fixed by https://github.com/ManageIQ/manageiq/pull/21623 Closing as duplicate of https://github.com/ManageIQ/manageiq-providers-vmware/issues/727

You can upgrade to a version which includes this fix, lasker branch has the fix but no release tag, so morphy-1 is the first release to include this fix

szajac01 commented 2 years ago

@agrare, we can't upgrade easily our platform so i have some further questions .

What is the impact of this bug if we leave like this ?
I tried to make the code change (https://github.com/ManageIQ/manageiq/pull/21623) on our kasparov-2 and now there's a refresh failure MIQ(ManageIQ::Providers::Vmware::InfraManager::Inventory::Collector#vim_collector) EMS: [xxxx], id: [4] Refresh failed [RuntimeError]: Invalid definition of index :manager_ref, there is no attribute :resource_name on model CustomAttribute Method:[block (2 levels) in ]

agrare commented 2 years ago

Hey @szajac01 it is %i(resource name) with a space not %i(resource_name) with an underscore. The impact is tags which are collected from the vSphere provider weren't added to all of the vms/hosts but rather only to a single one.

szajac01 commented 2 years ago

Hi @agrare , exact sorry, i fixed it. Don't have those message now but new ones... Warnings like WARN -- : MIQ(MiqEvent.raise_evm_event) Event CloneVM_Task_Complete for class [ManageIQ::Providers::Vmware::InfraManager::Vm] id [5231] was not raised: CloneVM_Task_Complete is not defined in MiqEventDefinition

WARN -- : MIQ(MiqEvent.raise_evm_event) Event PowerOnVM_Task_Complete for class [ManageIQ::Providers::Vmware::InfraManager::Vm] id [5642] was not raised: PowerOnVM_Task_Complete is not defined in MiqEventDefinition

agrare commented 2 years ago

@szajac01 that is an unrelated warning, is the refresh working for you now?