More robust handling of exceptions inside share manager ECS task

[SofiaSazonova]

Right now if an exception is raised inside approve_share of revoke_share job the status of the share will be stuck in "in_progress" state. What do we need to do:

1. Move share object status update to "finally" job
2. In case dataset_lock is failed to be released, we need to think about expiry mechanism
3. Dataset_lock can be a piece of information we may want to show in UI

[dlpzx]

Hi @SofiaSazonova I am working on the share managers a lot in #1283. This could be a part X in the list of items to improve the sharing. I have added it to the list of todos

[TejasRGitHub]

Along with the share manager, please check out this issue to add robust handling in ECS task for share verifier - #1300