Due to complicated access checks (reads and writes execute in their own access context) on some repositories (GCS, Azure, HDFS), using a hard coded buffer size of 4k for restores was needlessly inefficient.
By the same token, the use of stream copying with the default 8k buffer size for blob writes was inefficient as well.
We also had dedicated, undocumented buffer size settings for HDFS and FS repositories. For these two we would use a 100k buffer by default. We did not have such a setting for e.g. GCS though, which would only use an 8k read buffer which is needlessly small for reading from a raw `URLConnection`.
This commit adds an undocumented setting that sets the default buffer size to `128k` for all repositories. It removes wasteful allocation of such a large buffer for small writes and reads in case of HDFS and FS repositories (i.e. still using the smaller buffer to write metadata) but uses a large buffer for doing restores and uploading segment blobs.
This should speed up Azure and GCS restores and snapshots in a non-trivial way as well as save some memory when reading small blobs on FS and HFDS repositories.
In order to ensure that we do not write a broken piece of `RepositoryData`
because the phyiscal repository generation was moved ahead more than one step
by erroneous concurrent writing to a repository we must check whether or not
the current assumed repository generation exists in the repository physically.
Without this check we run the risk of writing on top of stale cached repository data.
Relates #56911
Restoring from a snapshot (which is a particular form of recovery) does not currently take recovery throttling into account
(i.e. the `indices.recovery.max_bytes_per_sec` setting). While restores are subject to their own throttling (repository
setting `max_restore_bytes_per_sec`), this repository setting does not allow for values to be configured differently on a
per-node basis. As restores are very similar in nature to peer recoveries (streaming bytes to the node), it makes sense to
configure throttling in a single place.
The `max_restore_bytes_per_sec` setting is also changed to default to unlimited now, whereas previously it was set to
`40mb`, which is the current default of `indices.recovery.max_bytes_per_sec`). This means that no behavioral change
will be observed by clusters where the recovery and restore settings were not adapted.
Relates https://github.com/elastic/elasticsearch/issues/57023
Co-authored-by: James Rodewig <james.rodewig@elastic.co>
This is a simple naming change PR, to fix the fact that "metadata" is a
single English word, and for too long we have not followed general
naming conventions for it. We are also not consistent about it, for
example, METADATA instead of META_DATA if we were trying to be
consistent with MetaData (although METADATA is correct when considered
in the context of "metadata"). This was a simple find and replace across
the code base, only taking a few minutes to fix this naming issue
forever.
* Remove BlobContainer Tests against Mocks
Removing all these weird mocks as asked for by #30424.
All these tests are now part of real repository ITs and otherwise left unchanged if they had
independent tests that didn't call the `createBlobStore` method previously.
The HDFS tests also get added coverage as a side-effect because they did not have an implementation
of the abstract repository ITs.
Closes#30424
* Remove Unused Single Delete in BlobStoreRepository
There are no more production uses of the non-bulk delete or the delete that throws
on missing so this commit removes both these methods.
Only the bulk delete logic remains. Where the bulk delete was derived from single deletes,
the single delete code was inlined into the bulk delete method.
Where single delete was used in tests it was replaced by bulk deleting.
* Make BlobStoreRepository Aware of ClusterState (#49639)
This is a preliminary to #49060.
It does not introduce any substantial behavior change to how the blob store repository
operates. What it does is to add all the infrastructure changes around passing the cluster service to the blob store, associated test changes and a best effort approach to tracking the latest repository generation on all nodes from cluster state updates. This brings a slight improvement to the consistency
by which non-master nodes (or master directly after a failover) will be able to determine the latest repository generation. It does not however do any tricky checks for the situation after a repository operation
(create, delete or cleanup) that could theoretically be used to get even greater accuracy to keep this change simple.
This change does not in any way alter the behavior of the blobstore repository other than adding a better "guess" for the value of the latest repo generation and is mainly intended to isolate the actual logical change to how the
repository operates in #49060
All the implementations of `EsBlobStoreTestCase` use the exact same
bootstrap code that is also used by their implementation of
`EsBlobStoreContainerTestCase`.
This means all tests might as well live under `EsBlobStoreContainerTestCase`
saving a lot of code duplication. Also, there was no HDFS implementation for
`EsBlobStoreTestCase` which is now automatically resolved by moving the tests over
since there is a HDFS implementation for the container tests.
This commit change the repositories base paths used in Azure/S3/GCS
integration tests so that they don't conflict with each other when tests
run in parallel on real storage services.
Closes#47202
This commit starts from the simple premise that the use of node settings
in blob store repositories is a mistake. Here we see that the node
settings are used to get default settings for store and restore throttle
rates. Yet, since there are not any node settings registered to this
effect, there can never be a default setting to fall back to there, and
so we always end up falling back to the default rate. Since this was the
only use of node settings in blob store repository, we move them. From
this, several places fall out where we were chaining settings through
only to get them to the blob store repository, so we clean these up as
well. That leaves us with the changeset in this commit.
* Repository Cleanup Endpoint (#43900)
* Snapshot cleanup functionality via transport/REST endpoint.
* Added all the infrastructure for this with the HLRC and node client
* Made use of it in tests and resolved relevant TODO
* Added new `Custom` CS element that tracks the cleanup logic.
Kept it similar to the delete and in progress classes and gave it
some (for now) redundant way of handling multiple cleanups but only allow one
* Use the exact same mechanism used by deletes to have the combination
of CS entry and increment in repository state ID provide some
concurrency safety (the initial approach of just an entry in the CS
was not enough, we must increment the repository state ID to be safe
against concurrent modifications, otherwise we run the risk of "cleaning up"
blobs that just got created without noticing)
* Isolated the logic to the transport action class as much as I could.
It's not ideal, but we don't need to keep any state and do the same
for other repository operations
(like getting the detailed snapshot shard status)
* We only use this method in one place in production code and can replace that with a read -> remove it to simplify the interface
* Keep it as an implementation detail in the Azure repository
Due to recent changes are done for converting `repository-hdfs` to test
clusters (#41252), the `integTestSecure*` tasks did not depend on
`secureHdfsFixture` which when running would fail as the fixture
would not be available. This commit adds the dependency of the fixture
to the task.
The `secureHdfsFixture` is a `AntFixture` which is spawned a process.
Internally it waits for 30 seconds for the resources to be made available.
For my local machine, it took almost 45 seconds to be available so I have
added the wait time as an input to the `AntFixture` defaults to 30 seconds
and set it to 60 seconds in case of secure hdfs fixture.
The integ test for secure hdfs was disabled for a long time and so
the changes done in #42090 to fix the tests are also done in this commit.
* Use ability to list child "folders" in the blob store to implement recursive delete on all stale index folders when cleaning up instead of using the diff between two `RepositoryData` instances to cover aborted deletes
* Runs after ever delete operation
* Relates #13159 (fixing most of this issues caused by unreferenced indices, leaving some meta files to be cleaned up only)
This is a prerequisite of #42189:
* Add directory delete method to blob container specific to each implementation:
* Some notes on the implementations:
* AWS + GCS: We can simply exploit the fact that both AWS and GCS return blobs lexicographically ordered which allows us to simply delete in the same order that we receive the blobs from the listing request. For AWS this simply required listing without the delimiter setting (so we get a deep listing) and for GCS the same behavior is achieved by not using the directory mode on the listing invocation. The nice thing about this is, that even for very large numbers of blobs the memory requirements are now capped nicely since we go page by page when deleting.
* For Azure I extended the parallelization to the listing calls as well and made it work recursively. I verified that this works with thread count `1` since we only block once in the initial thread and then fan out to a "graph" of child listeners that never block.
* HDFS and FS are trivial since we have directory delete methods available for them
* Enhances third party tests to ensure the new functionality works (I manually ran them for all cloud providers)
* Add Ability to List Child Containers to BlobContainer (#42653)
* Add Ability to List Child Containers to BlobContainer
* This is a prerequisite of #42189
* Remove Delete Method from BlobStore (#41619)
* The delete method on the blob store was used almost nowhere and just duplicates the delete method on the blob containers
* The fact that it provided for some recursive delete logic (that did not behave the same way on all implementations) was not used and not properly tested either
Motivated by slow snapshot deletes reported in e.g. #39656 and the fact that these likely are a contributing factor to repositories accumulating stale files over time when deletes fail to finish in time and are interrupted before they can complete.
* Makes snapshot deletion async and parallelizes some steps of the delete process that can be safely run concurrently via the snapshot thread poll
* I did not take the biggest potential speedup step here and parallelize the shard file deletion because that's probably better handled by moving to bulk deletes where possible (and can still be parallelized via the snapshot pool where it isn't). Also, I wanted to keep the size of the PR manageable.
* See https://github.com/elastic/elasticsearch/pull/39656#issuecomment-470492106
* Also, as a side effect this gives the `SnapshotResiliencyTests` a little more coverage for master failover scenarios (since parallel access to a blob store repository during deletes is now possible since a delete isn't a single task anymore).
* By adding a `ThreadPool` reference to the repository this also lays the groundwork to parallelizing shard snapshot uploads to improve the situation reported in #39657
Blob store compression was not enabled for some of the files in
snapshots due to constructor accessing sub-class fields. Fixed to
instead accept compress field as constructor param. Also fixed chunk
size validation to work.
Deprecated repositories.fs.compress setting as well to be able to unify
in a future commit.
This commit changes the `TransportVerifyShardBeforeCloseAction` so that it
always forces the flush of the shard. It seems that #37961 is not sufficient to
ensure that the translog and the Lucene commit share the exact same max
seq no and global checkpoint information in case of one or more noop
operations have been made.
The `BulkWithUpdatesIT.testThatMissingIndexDoesNotAbortFullBulkRequest`
and `FrozenIndexTests.testFreezeEmptyIndexWithTranslogOps` test this trivial
situation and they both fail 1 on 10 executions.
Relates to #33888
* There is no point in hsyncing after every individual write since there is only value in completely written blobs for restores, this is ensures by the `SYNC` flag already and there is no need for separately invoking `hsync` after individual writes
This commit changes the format of the `hits.total` in the search response to be an object with
a `value` and a `relation`. The `value` indicates the number of hits that match the query and the
`relation` indicates whether the number is accurate (in which case the relation is equals to `eq`)
or a lower bound of the total (in which case it is equals to `gte`).
This change also adds a parameter called `rest_total_hits_as_int` that can be used in the
search APIs to opt out from this change (retrieve the total hits as a number in the rest response).
Note that currently all search responses are accurate (`track_total_hits: true`) or they don't contain
`hits.total` (`track_total_hits: true`). We'll add a way to get a lower bound of the total hits in a
follow up (to allow numbers to be passed to `track_total_hits`).
Relates #33028
The contains syntax was added in #30874 but the skips were not properly
put in place.
The java runner has the feature so the tests will run as part of the
build, but language clients will be able to support it at their own
pace.
This removes custom Response classes that extend `AcknowledgedResponse` and do nothing, these classes are not needed and we can directly use the non-abstract super-class instead.
While this appears to be a large PR, no code has actually changed, only class names have been changed and entire classes removed.
Hadoop's security model uses the OS level authentication modules to collect
information about the current user. In JDK 11, the UnixLoginModule makes
use of a new permission to determine if the executing code is allowed to load
the libraries required to pull the user information from the OS. This PR adds
that permission and re-enables the tests that were previously failing when
testing against JDK 11.
In #29623 we added `Request` object flavored requests to the low level
REST client and in #30315 we deprecated the old `performRequest`s. This
changes most of the calls not in X-Pack to their new versions.
* Upgrade bouncycastle
Required to fix
`bcprov-jdk15on-1.55.jar; invalid manifest format `
on jdk 11
* Downgrade bouncycastle to avoid invalid manifest
* Add checksum for new jars
* Update tika permissions for jdk 11
* Mute test failing on jdk 11
* Add JDK11 to CI
* Thread#stop(Throwable) was removed
http://mail.openjdk.java.net/pipermail/core-libs-dev/2018-June/053536.html
* Disable failing tests #31456
* Temprorarily disable doc tests
To see if there are other failures on JDK11
* Only blacklist specific doc tests
* Disable only failing tests in ingest attachment plugin
* Mute failing HDFS tests #31498
* Mute failing lang-painless tests #31500
* Fix backwards compatability builds
Fix JAVA version to 10 for ES 6.3
* Add 6.x to bwx -> java10
* Prefix out and err from buildBwcVersion for readability
```
> Task :distribution:bwc:next-bugfix-snapshot:buildBwcVersion
[bwc] :buildSrc:compileJava
[bwc] WARNING: An illegal reflective access operation has occurred
[bwc] WARNING: Illegal reflective access by org.codehaus.groovy.reflection.CachedClass (file:/home/alpar/.gradle/wrapper/dists/gradle-4.5-all/cg9lyzfg3iwv6fa00os9gcgj4/gradle-4.5/lib/groovy-all-2.4.12.jar) to method java.lang.Object.finalize()
[bwc] WARNING: Please consider reporting this to the maintainers of org.codehaus.groovy.reflection.CachedClass
[bwc] WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations
[bwc] WARNING: All illegal access operations will be denied in a future release
[bwc] :buildSrc:compileGroovy
[bwc] :buildSrc:writeVersionProperties
[bwc] :buildSrc:processResources
[bwc] :buildSrc:classes
[bwc] :buildSrc:jar
```
* Also set RUNTIME_JAVA_HOME for bwcBuild
So that we can make sure it's not too new for the build to understand.
* Align bouncycastle dependency
* fix painles array tets
closes#31500
* Update jar checksums
* Keep 8/10 runtime/compile untill consensus builds on 11
* Only skip failing tests if running on Java 11
* Failures are dependent of compile java version not runtime
* Condition doc test exceptions on compiler java version as well
* Disable hdfs tests based on runtime java
* Set runtime java to minimum supported for bwc
* PR review
* Add comment with ticket for forbidden apis
Adds a new parameter to the BlobContainer#write*Blob methods to specify whether the existing file
should be overridden or not. For some metadata files in the repository, we actually want to replace
the current file. This is currently implemented through an explicit blob delete and then a fresh write.
In case of using a cloud provider (S3, GCS, Azure), this results in 2 API requests instead of just 1.
This change will therefore allow us to achieve the same functionality using less API requests.
* remove left-over comment
* make sure of the property for plugins
* skip installing modules if these exist in the distribution
* Log the distrbution being ran
* Don't allow running with integ-tests-zip passed externally
* top level x-pack/qa can't run with oss distro
* Add support for matching objects in lists
Makes it possible to have a key that points to a list and assert that a
certain object is present in the list. All keys have to be present and
values have to match. The objects in the source list may have additional
fields.
example:
```
match: { 'nodes.$master.plugins': { name: ingest-attachment } }
```
* Update plugin and module tests to work with other distributions
Some of the tests expected that the integration tests will always be ran
with the `integ-test-zip` distribution so that there will be no other
plugins loaded.
With this change, we check for the presence of the plugin without
assuming exclusivity.
* Allow modules to run on other distros as well
To match the behavior of tets.distributions
* Add and use a new `contains` assertion
Replaces the previus changes that caused `match` to do a partial match.
* Implement PR review comments
There's no need for an extra `blobExists()` call when writing a blob to the HDFS service. The writeBlob implementation for the HDFS repository already uses the `CreateFlag.CREATE` option on the file creation, which ensures that the blob that's uploaded does not already exist. This saves one network roundtrip.
* Sense HA HDFS settings and remove permission restrictions during regular execution.
This PR adds integration tests for HA-Enabled HDFS deployments, both regular and secured.
The Mini HDFS fixture has been updated to optionally run in HA-Mode. A new test suite has
been added for reproducing the effects of a Namenode failing over during regular repository
usage. Going forward, the HDFS Repository will still be subject to its self imposed permission
restrictions during normal use, but will no longer restrict them when running against an HA
enabled HDFS cluster. Instead, the plugin will rely on the provided security policy and not
further restrict the permissions so that the transparent operation to failover to a different
Namenode in the client does not raise security exceptions. Additionally, we are now testing the
secure mode with SASL based wire encryption of data between Elasticsearch and HDFS. This
includes a missing library (commons codec) in order to support this change.
For FsBlobStore and HdfsBlobStore, if the repository is read only, the blob store should be aware of the readonly setting and do not create directories if they don't exist.
Closes#21495
Since `#getAsMap` exposes internal representation we are trying to remove it
step by step. This commit is cleaning up some xcontent writing as well as
usage in tests
Add checks for special permissions before reading hdfs stream data. Also adds test from
readonly repository fix. MiniHDFS will now start with an existing repository with a single snapshot
contained within. Readonly Repository is created in tests and attempts to list the snapshots
within this repo.
Hadoop 2.7.x libraries fail when running on JDK9 due to the version string changing to a single
character. On Hadoop 2.8, this is no longer a problem, and it is unclear on whether the fix will be
backported to the 2.7 branch. This commit upgrades our dependency of Hadoop for the HDFS
Repository to 2.8.1.
This commit renames all rest test files to use the .yml extension
instead of .yaml. This way the extension used within all of
elasticsearch for yaml is consistent.
This PR introduces a subproject in test/fixtures that contains a Vagrantfile used for standing up a
KRB5 KDC (Kerberos). The PR also includes helper scripts for provisioning principals, a few
changes to the HDFS Fixture to allow it to interface with the KDC, as well as a new suite of
integration tests for the HDFS Repository plugin.
The HDFS Repository plugin senses if the local environment can support the HDFS Fixture
(Windows is generally a restricted environment). If it can use the regular fixture, it then tests if
Vagrant is installed with a compatible version to determine if the secure test fixtures should be
enabled. If the secure tests are enabled, then we create a Kerberos KDC fixture, tasks for adding
the required principals, and an HDFS fixture configured for security. A new integration test task is
also configured to use the KDC and secure HDFS fixture and to run a testing suite that uses
authentication. At the end of the secure integration test the fixtures are torn down.
Added missing permissions required for authenticating with Kerberos to HDFS. Also implemented
code to support authentication in the form of using a Kerberos keytab file. In order to support
HDFS authentication, users must install a Kerberos keytab file on each node and transfer it to the
configuration directory. When a user specifies a Kerberos principal in the repository settings the
plugin automatically enables security for Hadoop and begins the login process. There will be a
separate PR and commit for the testing infrastructure to support these changes.
This change simplifies how the rest test runner finds test files and
removes all leniency. Previously multiple prefixes and suffixes would
be tried, and tests could exist inside or outside of the classpath,
although outside of the classpath never quite worked. Now only classpath
tests are supported, and only one resource prefix is supported,
`/rest-api-spec/tests`.
closes#20240
We have a bunch of interfaces that have only a single implementation
for 6 years now. These interfaces are pretty useless from a SW development
perspective and only add unnecessary abstractions. They also require
lots of casting in many places where we expect that there is only one
concrete implementation. This change removes the interfaces, makes
all of the classes final and removes the duplicate `foo` `getFoo` accessors
in favor of `getFoo` from these classes.
This is related to #22116. Core no longer needs `SocketPermission`
`connect`.
This permission is relegated to these modules/plugins:
- transport-netty4 module
- reindex module
- repository-url module
- discovery-azure-classic plugin
- discovery-ec2 plugin
- discovery-gce plugin
- repository-azure plugin
- repository-gcs plugin
- repository-hdfs plugin
- repository-s3 plugin
And for tests:
- mocksocket jar
- rest client
- httpcore-nio jar
- httpasyncclient jar
Rory Hunter
Armin Braun
Yannick Welsch
Tanguy Leroux
Jason Tedor
Yogesh Gaikwad
Henning Andersen
Jim Ferenczi
Gordon Brown
Pratik Sanglikar
Alpar Torok
Benjamin Trent
Lee Hinman
James Baiera
Nik Everett
Vladimir Dolzhenko
kel
Simon Willnauer
Ryan Ernst
Tim Brooks