Adds the option fs.s3a.requester.pays.enabled, which, if set to true, allows
the client to access S3 buckets where the requester is billed for the IO.
Contributed by Daniel Carl Jones
* The source files for hdfs_ls
uses getopt for parsing the
command line arguments.
* getopt is available only on
Linux and thus, isn't cross
platform.
* Thus, we need to replace
getopt with
boost::program_options to
make this tool cross platform.
Optimize the scan for s3 by performing a deep tree listing,
inferring directory counts from the paths returned.
Contributed by Ahmar Suhail.
Change-Id: I26ffa8c6f65fd11c68a88d6e2243b0eac6ffd024
RBF proxy. There is a new configuration knob dfs.namenode.ip-proxy-users that configures
the list of users than can set their client ip address using the client context.
Fixes#4081
* The source files for hdfs_find uses
getopt for parsing the command
line arguments. getopt is available
only on Linux and thus, isn't cross
platform.
* Thus, we need to replace getopt
with boost::program_options to
make hdfs_find cross platform.
Follow-on patch to MAPREDUCE-7341, adding ABFS support and tests
* resilient rename
* tests for job commit through the manifest committer.
contains
- HADOOP-17976. ABFS etag extraction inconsistent between LIST and HEAD calls
- HADOOP-16204. ABFS tests to include terasort
Contributed by Steve Loughran.
Change-Id: I0a7d4043bdf19bcb00c033fc389730109b93b77f
This is a mapreduce/spark output committer optimized for
performance and correctness on Azure ADLS Gen 2 storage
(via the abfs connector) and Google Cloud Storage
(via the external gcs connector library).
* It is safe to use with HDFS, however it has not been optimized
for that use.
* It is *not* safe for use with S3, and will fail if an attempt
is made to do so.
Contributed by Steve Loughran
Change-Id: I6f3502e79c578b9fd1a8c1485f826784b5421fca
* New statistic names in StoreStatisticNames
(for joint use with s3a committers)
* Improvements to IOStatistics implementation classes
* RateLimiting wrapper to guava RateLimiter
* S3A committer Tasks moved over as TaskPool and
added support for RemoteIterator
* JsonSerialization.load() to fail fast if source does not exist
+ tests.
This commit is a prerequisite for the main MAPREDUCE-7341 Manifest Committer
patch.
Contributed by Steve Loughran
Change-Id: Ia92e2ab5083ac3d8d3d713a4d9cb3e9e0278f654
To get the new behavior, define fs.viewfs.trash.force-inside-mount-point to be true.
If the trash root for path p is in the same mount point as path p,
and one of:
* The mount point isn't at the top of the target fs.
* The resolved path of path is root (eg it is the fallback FS).
* The trash root isn't in user's target fs home directory.
get the corresponding viewFS path for the trash root and return it.
Otherwise, use <mnt>/.Trash/<user>.
Signed-off-by: Owen O'Malley <oomalley@linkedin.com>
Junfan Zhang
litao
He Xiaoqiao
Szilard Nemeth
Kengo Seki
zhongjingxiong
Akira Ajisaka
9uapaw
Lei Yang
Owen O'Malley
Masatake Iwasaki
GuoPhilipse
PJ Fanning
Tamas Domok
Benjamin Teke
Daniel Carl Jones
qinyuren
Gautham B A
André Fonseca
Steve Loughran
Szilard Nemeth
Abhishek Das
Viraj Jasani
Xing Lin