It is unclear for me, how the transactions implemented in google's datastore.
-
How they determines the resource I'm trying to reach? Because we are receiving global transaction object that we must use then for all requests. I have doubts, that this is done on the client library level, not on database level.
-
Is this transaction works both for read and write? I mean, that can I using the transaction implement document lock mechanism, so two user's cannot access it simultaneously (like mutex when we are dealing with multithreaded/multiprocessed applications).
-
And the final question. Does anybody know how the transnational mechanism implemented in datastore? I mean, high level architecture, or maybe processes diagrams, maybe short inside description, to get better understanding, what I am working with.
It shouldn't be relevant, but I'm using Google Cloud Functions and node.js environment for the project. I believe that this should not make any hard restrictions on usage.
via QuestionAndAnswer
No comments:
Post a Comment