add refresh after creating manager

[Lauferin]

Until now, when adding a storage manager sometimes it took a long time until an automatic refresh happened, and this caused that the user wasn't able to add the type system of a new physical storage until a manual refresh.

So a refresh was added directly after connecting to the storage manager.

[miq-bot]

Checked commit Autosde@91f2fda with ruby 2.6.10, rubocop 1.28.2, haml-lint 0.35.0, and yamllint
1 file checked, 0 offenses detected
Everything looks fine. 🍰

[agrare]

@Lauferin were you running with rails s + simulate_queue_worker or the full appliance when you saw refresh taking forever?

When a refresh worker is started (which will happen on a real appliance or podified install) a refresh is queued right away before processing any work: https://github.com/ManageIQ/manageiq/blob/master/app/models/manageiq/providers/base_manager/refresh_worker/runner.rb#L14

This should be within a few seconds of adding the provider.

[agrare]

This is not a great place for this, connect is called in a lot of places.

[Lauferin]

I paid attention to that, but I thought maybe for all those cases a refresh would fit well.

But if not, would it be preferable to do it 'after_save'?

[Lauferin]

Ok, I understand that after_save is not option (I tried and saw an infinite loop...). Where would you recommend it then? I don't see anywhere else in the file where this could fit.

[Lauferin]

@Lauferin were you running with rails s + simulate_queue_worker or the full appliance when you saw refresh taking forever?

When a refresh worker is started (which will happen on a real appliance or podified install) a refresh is queued right away before processing any work: https://github.com/ManageIQ/manageiq/blob/master/app/models/manageiq/providers/base_manager/refresh_worker/runner.rb#L14

This should be within a few seconds of adding the provider.

It happens with rails and queue worker and also when not stopping the server at all. It doesn't take forever, but it can take up to one minute approximately.

[Lauferin]

@Lauferin were you running with rails s + simulate_queue_worker or the full appliance when you saw refresh taking forever?
When a refresh worker is started (which will happen on a real appliance or podified install) a refresh is queued right away before processing any work: https://github.com/ManageIQ/manageiq/blob/master/app/models/manageiq/providers/base_manager/refresh_worker/runner.rb#L14
This should be within a few seconds of adding the provider.

It happens with rails and queue worker and also when not stopping the server at all. It doesn't take forever, but it can take up to one minute approximately.

I tried again in both ways and I insist, it usually takes more or less a minute... I'll be happy to show you on live if you want :-) (have you checked?)

[agrare]

@Lauferin so just so I understand, if you run on an appliance or with lib/workers/bin/run_single_worker.rb you do see an initial refresh being queued via https://github.com/ManageIQ/manageiq/blob/master/app/models/manageiq/providers/base_manager/refresh_worker/runner.rb#L14 ?

I tried running this locally with lib/workers/bin/run_single_worker.rb --ems-id=4 --role=ems_inventory ManageIQ::Providers::Autosde::StorageManager::RefreshWorker and I see an initial refresh queued right away on startup and it is run right away

[----] I, [2022-09-29T15:36:17.783156 #84125:97b8]  INFO -- evm: MIQ(ManageIQ::Providers::Autosde::StorageManager::RefreshWorker::Runner#do_before_work_loop) EMS [autosde.local] as [root] Queueing initial refresh for EMS 4.
[----] I, [2022-09-29T15:36:17.816920 #84125:97b8]  INFO -- evm: MIQ(MiqQueue.put) Message id: [119], Zone: [default], Role: [ems_inventory], Server: [], MiqTask id: [], Handler id: [], Ident: [ems_4], Target id: [], Instance id: [], Task id: [], Command: [EmsRefresh.refresh], Timeout: [7200], Priority: [100], State: [ready], Deliver On: [], Data: [1 bytes], Args: []
[----] I, [2022-09-29T15:36:17.828397 #84125:97b8]  INFO -- evm: MIQ(MiqQueue.get) Message id: [119], Zone: [default], Role: [ems_inventory], Server: [], MiqTask id: [], Handler id: [MiqWorker 32], Ident: [ems_4], Target id: [], Instance id: [], Task id: [], Command: [EmsRefresh.refresh], Timeout: [7200], Priority: [100], State: [dequeue], Deliver On: [], Data: [1 bytes], Args: [], Dequeued in: [0.016824348] seconds
[----] I, [2022-09-29T15:36:17.828615 #84125:97b8]  INFO -- evm: MIQ(MiqQueue#deliver) Message id: [119], Delivering...
[----] I, [2022-09-29T15:36:17.875350 #84125:97b8]  INFO -- evm: MIQ(ManageIQ::Providers::Autosde::StorageManager::Refresher#refresh) Refreshing all targets...

It might take the first refresh more than a minute to finish but once the worker is started the first refresh should be queued and started in seconds. If the issue is that the first refresh takes a bit to complete, queueing another isn't going to do anything to help that.