Retry policy for gcs calls

Hi, I have a question about the retry policy which is used with gcs.

In gcs.cc, I see that tiledb overrides the default gcs storage client’s retry policy, setting it to LimitedTimeRetryPolicy. My question is: why? Because according to the gcs c++ client docs (Google Cloud Storage C++ Client: Google Cloud Storage C++ Cl):

The library automatically retries requests that fail with transient errors, and follows the recommended practice to backoff between retries. […] The default policies are to continue retrying for up to 15 minutes, and to use truncated (at 5 minutes) exponential backoff, doubling the maximum backoff period between retries. By default the library retries all operations, even those that are not idempotent.

I’d rather use the recommended truncated exponential backoff policy, but I presume there were good reasons to override it?

@vincentschut apologies for the late reply here. I believe the override was made to mimic our AWS SDK retry policies at the time. There is not a specific or justifiable reason for this behavior over the default the GCC SDK provides. As such we’ll look at making the change for the default retry policy. Expect this to be done in the next couple of week.