[NBS] DestroyVolume operation of DiskRegistry based volumes hangs after receiving an error from SchemeShard

[sharpeye]

Relates to #2542

:BLOCKSTORE_SERVICE INFO: Deleting volume: diskId = "nonrepl-disk", destroyIfBroken = 0, sync = 0, fillGeneration = 0
:BLOCKSTORE_DISK_REGISTRY INFO: mark disk "nonrepl-disk" for cleanup

:BLOCKSTORE_VOLUME INFO: [72075186237009571] Stop Partition before volume destruction
:BLOCKSTORE_VOLUME INFO: [72075186237009571] Send poison pill to partition [53228:7503624403698463600:7007]
:BLOCKSTORE_VOLUME INFO: [72075186237009571] Partition [53228:7503624403698463600:7007] stopped

:BLOCKSTORE_SERVICE ERROR: Volume "nonrepl-disk": drop failed, error E_REJECTED ESchemeOpDropBlockStoreVolume failed with reason: Too many requests

:BLOCKSTORE_SERVER WARN: cloud/blockstore/libs/diagnostics/server_stats.cpp:549: [d:nonrepl-disk] [c:68223abd-4261-cc2a-d7a8-e5aa-93a777d1] DestroyVolume #550802513302528388 RESPONSE request rejected (execution: 3.901ms, postponed: 0, predicted: 0, backoff: 0, size: 0 B, unaligned: 0, error: E_REJECTED ESchemeOpDropBlockStoreVolume failed with reason: Too many requests)

:BLOCKSTORE_SERVER INFO: cloud/blockstore/libs/diagnostics/server_stats.cpp:317: [d:nonrepl-disk] [c:68223abd-4261-cc2a-d7a8-e5aa-93a777d1] DestroyVolume #14753985991817143554 REQUEST { Headers { ClientId: "68223abd-4261-cc2a-d7a8-e5aa-93a777d1" Timestamp: 1747073725885811 RequestId: 14753985991817143554 RequestTimeout: 20000 } DiskId: "nonrepl-disk" }

:BLOCKSTORE_SERVICE INFO: Deleting volume: diskId = "nonrepl-disk", destroyIfBroken = 0, sync = 0, fillGeneration = 0
:BLOCKSTORE_SERVICE ERROR: Volume "nonrepl-disk": unable to stat volume: E_REJECTED StatVolume request rejected

Before sending TEvModifyVolumeRequest, volume becomes a zombie. So all retries of DestroyVolume will fail with an error: E_REJECTED StatVolume request rejected