Details
-
Sub-task
-
Status: Resolved
-
Major
-
Resolution: Duplicate
-
2.8.0
-
None
-
None
Description
Noticed today that when connections are down, S3A retries on UnknownHostExceptions logging noisily in the process.
- it should be possible to define or customize retry policies for an FS instance (fail fast, exponential backoff, etc)
- we may want to explicitly have a fail-fast-if-offline retry policy, catching the common connectivity ones.
Testing will be fun here.
Attachments
Issue Links
- is duplicated by
-
HADOOP-13786 Add S3A committers for zero-rename commits to S3 endpoints
- Resolved
- is part of
-
HADOOP-13786 Add S3A committers for zero-rename commits to S3 endpoints
- Resolved
- is related to
-
HADOOP-14303 Review retry logic on all S3 SDK calls, implement where needed
- Resolved
-
HADOOP-14564 s3a test can hang in teardown with network problems
- Resolved