[fix][broker] Avoid blocking the PulsarAdmin callback thread on the post-unload load-report write

[merlimat]

Motivation

When a namespace isolation policy changes, ClustersBase.filterAndUnloadMatchedNamespaceAsync unloads the affected namespaces and then writes the load report so the cluster rebalances quickly. That write ran inside FutureUtil.waitForAll(unloadAsync...).thenAccept(...), which executes on the thread that completes the adminClient.namespaces().unloadAsync(...) futures — a PulsarAdmin async-HTTP-client callback thread.

LoadManager.writeLoadReportOnZookeeper(true) blocks there: ModularLoadManagerImpl.writeBrokerDataOnZooKeeper does lock.lock() plus a metadata-store updateValue(...).join(). So a brief metadata write blocked the admin-client callback thread.

Modifications

Run the load-report write via thenAcceptAsync(..., pulsar().getExecutor()) so it executes on the broker executor (where short blocking metadata writes are routine) instead of the admin-client callback thread. The REST response still waits for the write (the chained future completes after it); the write remains best-effort (failure is logged and swallowed, as before).

Verifying this change

Covered by AdminApi2Test.testIsolationPolicyUnloadsNSWith* (AllScope, ChangedScope1/2, NoneScope, PrimaryChanged, ScopeMissing; persistent + non-persistent) — 12/12 pass.

Does this pull request potentially affect one of the following parts:

If the box was checked, please highlight the changes

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment