shutdown upon parentchain code update

[brenzi]

closes #1633

code updates can cause extrinsics to fail unless we can update metadata and extrinsics AdditionalParams.
Our architecture currently doesn't allow to update these dynamically.

Shutting down the service gracefully (assuming automatic restart) is safe and clean and probably doesn't even cause additional downtime as we would need to pause TOP's anyway to avoid race conditions

caveat:

• this shutdown CAN come too late! If a TOP is executed while CodeUpdated is between InBlock and Finalized on the Parentchain, the extrinsic will fail. See shutdown upon parentchain UpdateAuthorized #1635 for a possible workaround and safely unshield using balances.transfer on parentchain #1257 for a proper solution

[clangenb]

Seems like a legit approach to me, given the caveats we are aware of that you mentioned.

[brenzi]

field test almost successful. Paseo had a runtime upgrade today:
https://paseo.subscan.io/block/3796856

the worker did shut down:

[L1Event:TargetA] CodeUpdated. Initiating service shutdown to allow clean restart
[L1Event:TargetA] Subscription terminated
....
[!] Sidechain block pruning loop has terminated
....
[!] Sidechain block production loop has terminated
....
[!] [Integritee] parentchain block syncing has terminated
......
[!] [TargetA] parentchain block syncing has terminated
....
[L1Event:Integritee] Subscription terminated
[!] waiting for 4 sensitive threads to shut down gracefully
[!] All threads stopped gracefully.

and restart:

aesm_service: warning: Turn to daemon. Use "--no-daemon" option to execute in foreground.
[2024-11-12T12:43:27.150Z INFO  integritee_service::config] Starting service in existing directory /opt/sidechain.
....
[Integritee] last synced parentchain block: 1520817
....
[>] DCAP setup: register QE collateral
...
[2024-11-12T12:43:33.189Z WARN  substrate_api_client::rpc::tungstenite_client::client] Expected subscription, but received an id response instead: Object {"error": Object {"code": Number(1014), "data": String("The transaction has too low priority to replace another transaction already in the pool."), "message": String("Priority is too low: (1353 vs 333)")}, "id": String("1"), "jsonrpc": String("2.0")}

from then on, the validateer was stuck (rpc did respond!) and needed a manual restart. See #1647