Differentiate from "EOF out of range/end of GET" from
"EOF channel problems" through
two different subclasses of EOFException and input streams to always
retry on http channel errors; out of range GET requests are not retried.
Currently an EOFException is always treated as a fail-fast call in read()
This allows for all existing external code catching EOFException to handle
both, but S3AInputStream to cleanly differentiate range errors (map to -1)
from channel errors (retry)
- HttpChannelEOFException is subclass of EOFException, so all code
which catches EOFException is still happy.
retry policy: connectivityFailure
- RangeNotSatisfiableEOFException is the subclass of EOFException
raised on 416 GET range errors.
retry policy: fail
- Method ErrorTranslation.maybeExtractChannelException() to create this
from shaded/unshaded NoHttpResponseException, using string match to
avoid classpath problems.
- And do this for SdkClientExceptions with OpenSSL error code WFOPENSSL0035.
We believe this is the OpenSSL equivalent.
- ErrorTranslation.maybeExtractIOException() to perform this translation as
appropriate.
S3AInputStream.reopen() code retries on EOF, except on
RangeNotSatisfiableEOFException,
which is converted to a -1 response to the caller
as is done historically.
S3AInputStream knows to handle these with
read(): HttpChannelEOFException: stream aborting close then retry
lazySeek(): Map RangeNotSatisfiableEOFException to -1, but do not map
any other EOFException class raised.
This means that
* out of range reads map to -1
* channel problems in reopen are retried
* channel problems in read() abort the failed http connection so it
isn't recycled
Tests for this using/abusing mocking.
Testing through actually raising 416 exceptions and verifying that
readFully(), char read() and vector reads are all good.
There is no attempt to recover within a readFully(); there's
a boolean constant switch to turn this on, but if anyone does
it a test will spin forever as the inner PositionedReadable.read(position, buffer, len)
downgrades all EOF exceptions to -1.
A new method would need to be added which controls whether to downgrade/rethrow
exceptions.
What does that mean? Possibly reduced resilience to non-retried failures
on the inner stream, even though more channel exceptions are retried on.
Contributed by Steve Loughran
Move to the new auth flow based signers for aws. * Implement a new Signer Initialization Chain
* Add a new instantiation method
* Add a new test
* Fix Reflection Code for SignerInitialization
Contributed by Harshit Gupta
The option fs.azure.footer.read.request.size sets the size of the footer to
read and cache; the default value of 524288 has been measured to
be good for most workloads running on parquet, ORC and similar file formats.
Contributed by Anuj Modi
Adds a new option fs.s3a.s3express.create.session; default is true.
When false, this disables the CreateSession call to create/refresh temporary
session credentials when working with an Amazon S3 Express store.
This avoids having to give the caller the new IAM role permission,
at the expense of every remote call on the S3 Express store having to
include the latency of a checkup of IAM permissions.
* fs.s3a.s3express.create.session is set to false in tests which generate new
role permissions and call AssumeRole
* move ApiCallTimeoutException logic until after sdk exceptions get translated
to IOE. This lines up for any future AWS throwing up underlying cause here.
* Tests will automatically skip ACL, storage class, S3 Select or encryption tests
when target fs is S3Express.
* same for the out of order multipart uploader test cases, v1 listing.
* bucket tool s3 test treats invalid location error as a successful invocation
of the create bucket attempt
Contributed by Steve Loughran
This adds borad support for Amazon S3 Express One Zone to the S3A connector,
particularly resilience of other parts of the codebase to LIST operations returning
paths under which only in-progress uploads are taking place.
hadoop-common and hadoop-mapreduce treewalking routines all cope with this;
distcp is left alone.
There are still some outstanding followup issues, and we expect more to surface
with extended use.
Contains HADOOP-18955. AWS SDK v2: add path capability probe "fs.s3a.capability.aws.v2
* lets us probe for AWS SDK version
* bucket-info reports it
Contains HADOOP-18961 S3A: add s3guard command "bucket"
hadoop s3guard bucket -create -region us-west-2 -zone usw2-az2 \
s3a://stevel--usw2-az2--x-s3/
* requires -zone if bucket is zonal
* rejects it if not
* rejects zonal bucket suffixes if endpoint is not aws (safety feature)
* imperfect, but a functional starting point.
New path capability "fs.s3a.capability.zonal.storage"
* Used in tests to determine whether pending uploads manifest paths
* cli tests can probe for this
* bucket-info reports it
* some tests disable/change assertions as appropriate
----
Shell commands fail on S3Express buckets if pending uploads.
New path capability in hadoop-common
"fs.capability.directory.listing.inconsistent"
1. S3AFS returns true on a S3 Express bucket
2. FileUtil.maybeIgnoreMissingDirectory(fs, path, fnfe)
decides whether to swallow the exception or not.
3. This is used in: Shell, FileInputFormat, LocatedFileStatusFetcher
Fixes with tests
* fs -ls -R
* fs -du
* fs -df
* fs -find
* S3AFS.getContentSummary() (maybe...should discuss)
* mapred LocatedFileStatusFetcher
* Globber, HADOOP-15478 already fixed that when dealing with
S3 inconsistencies
* FileInputFormat
S3Express CreateSession request is permitted outside audit spans.
S3 Bulk Delete calls request the store to return the list of deleted objects
if RequestFactoryImpl is set to trace.
log4j.logger.org.apache.hadoop.fs.s3a.impl.RequestFactoryImpl=TRACE
Test Changes
* ITestS3AMiscOperations removes all tests which require unencrypted
buckets. AWS S3 defaults to SSE-S3 everywhere.
* ITestBucketTool to test new tool without actually creating new
buckets.
* S3ATestUtils add methods to skip test suites/cases if store is/is not
S3Express
* Cutting down on "is this a S3Express bucket" logic to trailing --x-s3 string
and not worrying about AZ naming logic. commented out relevant tests.
* ITestTreewalkProblems validated against standard and S3Express stores
Outstanding
* Distcp: tests show it fails. Proposed: release notes.
---
x-amz-checksum header not found when signing S3Express messages
This modifies the custom signer in ITestCustomSigner to be a subclass
of AwsS3V4Signer with a goal of preventing signing problems with
S3 Express stores.
----
RemoteFileChanged renaming multipart file
Maps 412 status code to RemoteFileChangedException
Modifies huge file tests
-Adds a check on etag match for stat vs list
-ITestS3AHugeFilesByteBufferBlocks renames parent dirs, rather than
files, to replicate distcp better.
----
S3Express custom Signing cannot handle bulk delete
Copy custom signer into production JAR, so enable downstream testing
Extend ITestCustomSigner to cover more filesystem operations
- PUT
- POST
- COPY
- LIST
- Bulk delete through delete() and rename()
- list + abort multipart uploads
Suite is parameterized on bulk delete enabled/disabled.
To use the new signer for a full test run:
<property>
<name>fs.s3a.custom.signers</name>
<value>CustomSdkSigner:org.apache.hadoop.fs.s3a.auth.CustomSdkSigner</value>
</property>
<property>
<name>fs.s3a.s3.signing-algorithm</name>
<value>CustomSdkSigner</value>
</property>
Increases existing pool sizes, as with server scale and vector
IO, larger pools are needed
fs.s3a.connection.maximum 200
fs.s3a.threads.max 96
Adds new configuration options for v2 sdk internal timeouts,
both with default of 60s:
fs.s3a.connection.acquisition.timeout
fs.s3a.connection.idle.time
All the pool/timoeut options are covered in performance.md
Moves all timeout/duration options in the s3a FS to taking
temporal units (h, m, s, ms,...); retaining the previous default
unit (normally millisecond)
Adds a minimum duration for most of these, in order to recover from
deployments where a timeout has been set on the assumption the unit
was seconds, not millis.
Uses java.time.Duration throughout the codebase;
retaining the older numeric constants in
org.apache.hadoop.fs.s3a.Constants for backwards compatibility;
these are now deprecated.
Adds new class AWSApiCallTimeoutException to be raised on
sdk-related methods and also gateway timeouts. This is a subclass
of org.apache.hadoop.net.ConnectTimeoutException to support
existing retry logic.
+ reverted default value of fs.s3a.create.performance to false;
inadvertently set to true during testing.
Contributed by Steve Loughran.
Add a new option:
fs.s3a.optimized.copy.from.local.enabled
This will enable (default) or disable the
optimized CopyFromLocalOperation upload operation
when copyFromLocalFile() is invoked.
When false the superclass implementation is used; duration
statistics are still collected, though audit span entries
in logs will be for the individual fs operations, not the
overall operation.
Contributed by Steve Loughran
If fs.s3a.create.performance is set on a bucket
- All file overwrite checks are skipped, even if the caller says
otherwise.
- All directory existence checks are skipped.
- Marker deletion is *always* skipped.
This eliminates a HEAD and a LIST for every creation.
* New path capability "fs.s3a.create.performance.enabled" true
if the option is enabled.
* Parameterize ITestS3AContractCreate to expect the different
outcomes
* Parameterize ITestCreateFileCost similarly, with
changed cost assertions there.
* create(/) raises an IOE. existing bug only noticed here.
Contributed by Steve Loughran
S3A directory delete and rename will optionally abort all pending multipart uploads
in their under their to-be-deleted paths when.
fs.s3a.directory.operations.purge.upload is true
It is off by default.
The filesystems hasPathCapability("fs.s3a.directory.operations.purge.upload")
probe will return true when this feature is enabled.
Multipart uploads may accrue from interrupted data writes, uncommitted
staging/magic committer jobs and other operations/applications. On AWS S3
lifecycle rules are the recommended way to clean these; this change improves
support for stores which lack these rules.
Contributed by Steve Loughran
This restores asynchronous retrieval/refresh of any AWS credentials provided by the
EC2 instance/container in which the process is running.
Contributed by Steve Loughran
S3A region logic improved for better inference and
to be compatible with previous releases
1. If you are using an AWS S3 AccessPoint, its region is determined
from the ARN itself.
2. If fs.s3a.endpoint.region is set and non-empty, it is used.
3. If fs.s3a.endpoint is an s3.*.amazonaws.com url,
the region is determined by by parsing the URL
Note: vpce endpoints are not handled by this.
4. If fs.s3a.endpoint.region==null, and none could be determined
from the endpoint, use us-east-2 as default.
5. If fs.s3a.endpoint.region=="" then it is handed off to
The default AWS SDK resolution process.
Consult the AWS SDK documentation for the details on its resolution
process, knowing that it is complicated and may use environment variables,
entries in ~/.aws/config, IAM instance information within
EC2 deployments and possibly even JSON resources on the classpath.
Put differently: it is somewhat brittle across deployments.
Contributed by Ahmar Suhail
Tune AWS v2 SDK changes based on testing with third party stores
including GCS.
Contains HADOOP-18889. S3A v2 SDK error translations and troubleshooting docs
* Changes needed to work with multiple third party stores
* New third_party_stores document on how to bind to and test
third party stores, including google gcs (which works!)
* Troubleshooting docs mostly updated for v2 SDK
Exception translation/resilience
* New AWSUnsupportedFeatureException for unsupported/unavailable errors
* Handle 501 method unimplemented as one of these
* Error codes > 500 mapped to the AWSStatus500Exception if no explicit
handler.
* Precondition errors handled a bit better
* GCS throttle exception also recognized.
* GCS raises 404 on a delete of a file which doesn't exist: swallow it.
* Error translation uses reflection to create IOE of the right type.
All IOEs at the bottom of an AWS stack chain are regenerated.
then a new exception of that specific type is created, with the top level ex
its cause. This is done to retain the whole stack chain.
* Reduce the number of retries within the AWS SDK
* And those of s3a code.
* S3ARetryPolicy explicitly declare SocketException as connectivity failure
but subclasses BindException
* SocketTimeoutException also considered connectivity
* Log at debug whenever retry policies looked up
* Reorder exceptions to alphabetical order, with commentary
* Review use of the Invoke.retry() method
The reduction in retries is because its clear when you try to create a bucket
which doesn't resolve that the time for even an UnknownHostException to
eventually fail over 90s, which then hit the s3a retry code.
- Reducing the SDK retries means these escalate to our code better.
- Cutting back on our own retries makes it a bit more responsive for most real
deployments.
- maybeTranslateNetworkException() and s3a retry policy means that
unknown host exception is recognised and fails fast.
Contributed by Steve Loughran
The default value for fs.azure.data.blocks.buffer is changed from "disk" to "bytebuffer"
This will speed up writing to azure storage, at the risk of running out of memory
-especially if there are many threads writing to abfs at the same time and the
upload bandwidth is limited.
If jobs do run out of memory writing to abfs, change the option back to "disk"
Contributed by Anmol Asrani
Jobs which commit their work to S3 thr
magic committer now use a unique magic
containing the job ID:
__magic_job-${jobid}
This allows for multiple jobs to write
to the same destination simultaneously.
Contributed by Syed Shameerur Rahman
* The multipart flag fs.s3a.multipart.uploads.enabled is passed to the async client created
* s3A connector bypasses the transfer manager entirely if disabled or for small files.
Contributed by Steve Loughran
All uses of jersey-json in the yarn and other hadoop modules now
exclude the obsolete org.codehaus.jettison/jettison and so avoid
all security issues which can come from the library.
Contributed by PJ Fanning
This patch migrates the S3A connector to use the V2 AWS SDK.
This is a significant change at the source code level.
Any applications using the internal extension/override points in
the filesystem connector are likely to break.
This includes but is not limited to:
- Code invoking methods on the S3AFileSystem class
which used classes from the V1 SDK.
- The ability to define the factory for the `AmazonS3` client, and
to retrieve it from the S3AFileSystem. There is a new factory
API and a special interface S3AInternals to access a limited
set of internal classes and operations.
- Delegation token and auditing extensions.
- Classes trying to integrate with the AWS SDK.
All standard V1 credential providers listed in the option
fs.s3a.aws.credentials.provider will be automatically remapped to their
V2 equivalent.
Other V1 Credential Providers are supported, but only if the V1 SDK is
added back to the classpath.
The SDK Signing plugin has changed; all v1 signers are incompatible.
There is no support for the S3 "v2" signing algorithm.
Finally, the aws-sdk-bundle JAR has been replaced by the shaded V2
equivalent, "bundle.jar", which is now exported by the hadoop-aws module.
Consult the document aws_sdk_upgrade for the full details.
Contributed by Ahmar Suhail + some bits by Steve Loughran
Adds a class DelegationBindingInfo which contains binding info
beyond just the AWS credential list.
The binding class can be expanded when needed. Until then, all existing
implementations will work, as the new method
DelegationBindingInfo deploy(AbstractS3ATokenIdentifier retrievedIdentifier)
falls back to the original methods.
To avoid the ABFS instance getting closed due to GC while the streams are working, attach the ABFS instance to a backReference opaque object and passing down to the streams so that we have a hard reference while the streams are working.
Contributed by: Mehakmeet Singh
This modifies the manifest committer so that the list of files
to rename is passed between stages as a file of
writeable entries on the local filesystem.
The map of directories to create is still passed in memory;
this map is built across all tasks, so even if many tasks
created files, if they all write into the same set of directories
the memory needed is O(directories) with the
task count not a factor.
The _SUCCESS file reports on heap size through gauges.
This should give a warning if there are problems.
Contributed by Steve Loughran
This
1. changes the default value of fs.s3a.directory.marker.retention
to "keep"
2. no longer prints a message when an S3A FS instances is
instantiated with any option other than delete.
Switching to marker retention improves performance
on any S3 bucket as there are no needless marker DELETE requests
-leading to a reduction in write IOPS and and any delays waiting
for the DELETE call to finish.
There are *very* significant improvements on versioned buckets,
where tombstone markers slow down LIST operations: the more
tombstones there are, the worse query planning gets.
Having versioning enabled on production stores is the foundation
of any data protection strategy, so this has tangible benefits
in production.
It is *not* compatible with older hadoop releases; specifically
- Hadoop branch 2 < 2.10.2
- Any release of Hadoop 3.0.x and Hadoop 3.1.x
- Hadoop 3.2.0 and 3.2.1
- Hadoop 3.3.0
Incompatible releases have no problems reading data in stores
where markers are retained, but can get confused when deleting
or renaming directories.
If you are still using older versions to write to data, and cannot
yet upgrade, switch the option back to "delete"
Contributed by Steve Loughran
This:
1. Adds optLong, optDouble, mustLong and mustDouble
methods to the FSBuilder interface to let callers explicitly
passin long and double arguments.
2. The opt() and must() builder calls which take float/double values
now only set long values instead, so as to avoid problems
related to overloaded methods resulting in a ".0" being appended
to a long value.
3. All of the relevant opt/must calls in the hadoop codebase move to
the new methods
4. And the s3a code is resilient to parse errors in is numeric options
-it will downgrade to the default.
This is nominally incompatible, but the floating-point builder methods
were never used: nothing currently expects floating point numbers.
For anyone who wants to safely set numeric builder options across all compatible
releases, convert the number to a string and then use the opt(String, String)
and must(String, String) methods.
Contributed by Steve Loughran
Explicitly sets the fs.s3a.endpoint.region to eu-west-1 so
the ARN-referenced fs creation fails with unknown store
rather than IllegalArgumentException.
Steve Loughran
To support recovery of network failures during rename, the abfs client
fetches the etag of the source file, and when recovering from a
failure, uses this tag to determine whether the rename succeeded
before the failure happened.
* This works for files, but not directories
* It adds the overhead of a HEAD request before each rename.
* The option can be disabled by setting "fs.azure.enable.rename.resilience"
to false
Contributed by Sree Bhattacharyya
This change lets the client react pre-emptively to server load without getting to 503 and the exponential backoff
which follows. This stops performance suffering so much as capacity limits are approached for an account.
Contributed by Anmol Asranii
POM and LICENSE fixup of transient dependencies
* Exclude hadoop-cloud-storage imports which come in with hadoop-common
* Add explicit import of hadoop's org.codehaus.jettison declaration
to hadoop-aliyun
* Tune aliyun jars imports
* Update LICENSE-binary for the current set of libraries.
Contributed by Steve Loughran
Followup to the original HADOOP-18582.
Temporary path cleanup is re-enabled for -append jobs
as these will create temporary files when creating or overwriting files.
Contributed by Ayush Saxena
Adding toggleable support for modification time during distcp -update between two stores with incompatible checksum comparison.
Contributed by: Mehakmeet Singh <mehakmeet.singh.behl@gmail.com>
Fixes a javadoc error which came with
HADOOP-18577. ABFS: Add probes of readahead fix (#5205)
Part of the HADOOP-18521 ABFS readahead fix; MUST be included.
Contributed by Steve Loughran
Followup patch to HADOOP-18456 as part of HADOOP-18521,
ABFS ReadBufferManager buffer sharing across concurrent HTTP requests
Add probes of readahead fix aid in checking safety of
hadoop ABFS client across different releases.
* ReadBufferManager constructor logs the fact it is safe at TRACE
* AbfsInputStream declares it is fixed in toString()
by including fs.azure.capability.readahead.safe" in the
result.
The ABFS FileSystem hasPathCapability("fs.azure.capability.readahead.safe")
probe returns true to indicate the client's readahead manager has been fixed
to be safe when prefetching.
All Hadoop releases for which probe this returns false
and for which the probe "fs.capability.etags.available"
returns true at risk of returning invalid data when reading
ADLS Gen2/Azure storage data.
Contributed by Steve Loughran.
This has triggered an OOM in a process which was churning through s3a fs
instances; the increased memory footprint of IOStatistics amplified what
must have been a long-standing issue with FS instances being created
and not closed()
* Makes sure instrumentation is closed when the FS is closed.
* Uses a weak reference from metrics to instrumentation, so even
if the FS wasn't closed (see HADOOP-18478), this back reference
would not cause the S3AInstrumentation reference to be retained.
* If S3AFileSystem is configured to log at TRACE it will log the
calling stack of initialize(), so help identify where the
instance is being created. This should help track down
the cause of instance leakage.
Contributed by Steve Loughran.
This is a followup to the original HADOOP-18546
patch; cherry-picks of that should include this
or follow up with it.
Removes risk of race conditions in assertions
of ITestReadBufferManager on the state of the in-progress
and completed queues by removing assertions brittle
to race conditions in scheduling/network IO
* Waits for all the executor pool shutdown to complete before
making any assertions
* Assertions that there are no in progress reads MUST be
cut as there may be some and they won't be cancelled.
* Assertions that the completed list is without buffers
of a closed stream are brittle because if there was
an in progress stream which completed after stream.close()
then it will end up in the list.
Contributed by Steve Loughran
This addresses HADOOP-18521, "ABFS ReadBufferManager buffer sharing
across concurrent HTTP requests" by not trying to cancel
in progress reads.
It supercedes HADOOP-18528, which disables the prefetching.
If that patch is applied *after* this one, prefetching
will be disabled.
As well as changing the default value in the code,
core-default.xml is updated to set
fs.azure.enable.readahead = true
As a result, if Configuration.get("fs.azure.enable.readahead")
returns a non-null value, then it can be inferred that
it was set in or core-default.xml (the fix is present)
or in core-site.xml (someone asked for it).
Contributed by Pranav Saxena.
This allows abfs request throttling to be shared across all
abfs connections talking to containers belonging to the same abfs storage
account -as that is the level at which IO throttling is applied.
The option is enabled/disabled in the configuration option
"fs.azure.account.throttling.enabled";
The default is "true"
Contributed by Anmol Asrani
