[VL] Add support for GCS retry properties (#5858)

The Velox GCS connector was recently added a couple of properties:

HiveConfig::kGCSMaxRetryCount: Retry until a number of transient errors is detected.
HiveConfig::kGCSMaxRetryTime: Retry until a time elapses.
This properties are useful to allow a client to either fail immediately or keep retring for several minutes.

If none of them is set, then it will use the defaults set by GCS.

This change was tested manually, by entering an unnaccesible GCS bucket then verifying that the properties are honored.

cpp/velox/utils/ConfigExtractor.cc

@@ -121,6 +121,19 @@ std::shared_ptr<facebook::velox::core::MemConfig> getHiveConfig(std::shared_ptr<
     }
   }
 
+  // https://github.com/GoogleCloudDataproc/hadoop-connectors/blob/master/gcs/CONFIGURATION.md#http-transport-configuration
+  // https://cloud.google.com/cpp/docs/reference/storage/latest/classgoogle_1_1cloud_1_1storage_1_1LimitedErrorCountRetryPolicy
+  auto gsMaxRetryCount = conf->get("spark.hadoop.fs.gs.http.max.retry");
+  if (gsMaxRetryCount.hasValue()) {
+    hiveConfMap[facebook::velox::connector::hive::HiveConfig::kGCSMaxRetryCount] = gsMaxRetryCount.value();
+  }
+
+  // https://cloud.google.com/cpp/docs/reference/storage/latest/classgoogle_1_1cloud_1_1storage_1_1LimitedTimeRetryPolicy
+  auto gsMaxRetryTime = conf->get("spark.hadoop.fs.gs.http.max.retry-time");
+  if (gsMaxRetryTime.hasValue()) {
+    hiveConfMap[facebook::velox::connector::hive::HiveConfig::kGCSMaxRetryTime] = gsMaxRetryTime.value();
+  }
+
   // https://github.com/GoogleCloudDataproc/hadoop-connectors/blob/master/gcs/CONFIGURATION.md#authentication
   auto gsAuthType = conf->get("spark.hadoop.fs.gs.auth.type");
   if (gsAuthType.hasValue()) {

docs/get-started/VeloxGCS.md

@@ -36,4 +36,18 @@ spark.hadoop.fs.gs.auth.service.account.json.keyfile // path to the json file wi
 For cases when a GCS mock is used, an optional endpoint can be provided:
 ```sh
 spark.hadoop.fs.gs.storage.root.url  // url to the mock gcs service including starting with http or https
+```
+
+## Configuring GCS max retry count
+
+For cases when a transient server error is detected, GCS can be configured to keep retrying until a number of transient error is detected.
+```sh
+spark.hadoop.fs.gs.http.max.retry // number of times to keep retrying unless a non-transient error is detected
+```
+
+## Configuring GCS max retry time
+
+For cases when a transient server error is detected, GCS can be configured to keep retrying until the retry loop exceeds a prescribed duration.
+```sh
+spark.hadoop.fs.gs.http.max.retry-time // a string representing the time keep retring (10s, 1m, etc).
 ```