IGNITE-12701 : Disallow silent deactivation in CLI and REST.

[Vladsz83]

Changes in public behavior:

1. Involved optional flag 'force' for control.sh (bat) with --deactivate and --set-sate:
   control.sh --deactivate --force
   control.sh --set-state INACTIVE --force

2. Involved optional parameter 'force=true|false' for REST cluster state requests:
   http://127.0.0.1:8092/ignite?cmd=deactivate&force=true
   http://127.0.0.1:8092/ignite?cmd=inactive&force=true
   http://127.0.0.1:8092/ignite?cmd=setstate&state=INACTIVE&force=true

Major code changes:

1. Added GridClientClusterStateRequestV2.
2. Deprecated GridClientClusterStateRequest.
3. Added flag ‘force’ for REST in GridRestClusterStateRequest
4. Added flag ‘force’ CLI in DeactivateCommand, ClusterStateChangeCommand.
5. Added IgniteFeatures# FORCED_CHANGE_OF_CLUSTER_STATE.
6. Added checking of deactivation safety: GridClusterStateProcessor# isDeactivationSafe()
7. Involved o deactivation safety checking in GridClusterStateProcessing and GridClientClusterStateImpl.

[nizhikov]

It's not clear why we force deactivation here.

[Vladsz83]

It's not clear why we force deactivation here.

All previous internal changes of cluster state were forced. Thus, all current internal changes of cluster are considered forced. Only the exceptions are the user commands being translated to not-forced internal calls without flag ‘force’ passed.

[sk0x50]

Hello @nizhikov ,

Please don't get me wrong, but...
What is the reason to include all intermediate commit messages into the final one? 4921fcf
Moreover, it contains the following tickets: IGNITE-12597, IGNITE-12646, and IGNITE-12703 that not related to the issue at all. It looks like it was done by accident.

[Vladsz83]

Moreover, it contains the following tickets: IGNITE-12597, IGNITE-12646, and IGNITE-12703 that not related to the issue at all. It looks like it was done by accident.

Hi, @sk0x50

This merge doesn't contain mentioned issues IGNITE-12597, IGNITE-12646, and IGNITE-12703 at all. That was just wrong comments in my branch.

[sk0x50]

Hello @Vladsz83 ,

This merge doesn't contain mentioned issues IGNITE-12597, IGNITE-12646, and IGNITE-12703 at all. That was just wrong comments in my branch.

Yes, I know that, and you are right.
Perhaps, my message was not quite correct. I just wanted to say that it contains incorrect commit messages that can lead to a misunderstanding. That's all.

Thanks,
Slava.

[nizhikov]

Hello @sk0x50

First of all, this is my fault, but the reason for the fault is GitHub bug.
I wrote a good commit message before the merge.

But after pushing the "merge" button I've got "We can't merge your PR" error message and "Try again" button. After pushing "Try again" we got what we got - merge with the default commit message.

Sorry, for that.

[sk0x50]

Hello @nizhikov,

Ok, got it! Thank you for the clarification.

Thanks,
S.