Simple object with lock and expiration time not removed after locks are expired

[abereziny]

Let say we have:

1. Current epoch is 100
2. Simple object with __NEOFS__EXPIRATION_EPOCH=101
3. Simple object is locked with lock until epoch 103

Steps:
Tick epochs until current epoch = 103
Check object availability

Expected Behavior

Simple object should be removed since locks are gone and object expired

Current Behavior

Simple object continues to present in storage even if we wait additional extra epochs

Tests

in branch feature-abereziny-add-lock-tests-2

• pytest_tests/testsuites/object/test_object_lock.py#test_expired_object_should_be_deleted_after_locks_are_expired

[acid-ant]

Blocked by:

• Make it clear that EXPIRATION_EPOCH is the last "available" epoch for all entities neofs-api#253
• EXPIRATION_EPOCH must be the last "available" epoch for all entities neofs-sdk-go#370

[acid-ant]

it works as expected. Fixed help message for parameter here #2169. I've added one extra epoch tick and test passed.

[abereziny]

Currently it's not working as expected because objects got removed despite of locks

[acid-ant]

@abereziny it works as expected. It is ok to get not found error in your case, because expiration epoch in lock blocks any action with data on the file system, but do not affect presence of the object in system. Now it is impossible to test.
In other words, it is possible create object with expiration epoch N and lock with M where M > N. Object will be unavailable at N+1 and persists in fs until M+1

[abereziny]

@acid-ant I do not agree with that because current behavior have no real practical use case.
Why do we need to lock object which will be unavailable? What's the point?

for example: we have document with expiration, court need this document in the process so it must not be deleted and must be available until process is complete (despite it's expiration), so lock is assigned. If the document is not available - it practically means that lock doesn't work from user point of view.

[abereziny]

Logged and issue for API to further elaboration

[roman-khimov]

@evgeniiz321, do we have any test scenario like this?

[evgeniiz321]

@roman-khimov yes, we do - https://rest.fs.neo.org/HXSaMJXk2g8C14ht8HSi7BBaiYZ1HeWh2xnWPGQCg4H6/1784-1722447068/index.html#suites/295ed2e000f8fa6f3ade59cc13b98615/9a75854013e7509b/