Almost every outbound message is serialized to buffers of 16k pagesize.
We were serializing these messages off the IO loop (and retaining the concrete message
instance as well) and would then enqueue it on the IO loop to be dealt with as soon as the
channel is ready.
1. This would cause buffers to be held onto for longer than necessary, causing less reuse on average.
2. If a channel was slow for some reason, not only would concrete message instances queue up for it, but also 16k of buffers would be reserved for each message until it would be written+flushed physically.
With this change, the serialization happens on the event loop which effectively limits the number of buffers that `N` IO-threads will ever use so long as messages are small and channels writable.
Also, this change dereferences the reference to the concrete outbound message as soon as it has been serialized to save some more on GC.
This reduces the GC time for a default PMC run by about 50% in experiments (3 nodes, 2G heap each, loopback ... obvious caveat is that GC isn't that heavy in the first place with recent changes but still a measurable gain).
I also expect it to be helpful for master node stability by causing less of a spike if master is e.g. hit by a large number of requests that are processed batched (e.g. shard snapshot status updates) and responded to in a short time frame all at once.
Obviously, the downside to this change is that it introduces more latency on the IO loop for the serialization. But since we read all of these messages on the IO loop as well I don't see it as much of a qualitative change really and the more predictable buffer use seems much more valuable relatively.
We don't need to hold on to the request body past the beginning of sending
the response. There is no need to keep a reference to it until after the response
has been sent fully and we can eagerly release it here.
Note, this can be optimized further to release the contents even earlier but for now
this is an easy increment to saving some memory on the IO pool.
Elasticsearch requires that a HttpRequest abstraction be implemented
by http modules before server processing. This abstraction controls when
underlying resources are released. This commit moves this abstraction to
be created immediately after content aggregation. This change will
enable follow-up work including moving Cors logic into the server
package and tracking bytes as they are aggregated from the network
level.
In most cases we are seeing a `PooledHeapByteBuf` here now. No need to
redundantly create an new `ByteBuffer` and single element array for it
here when we can just directly unwrap its internal `byte[]`.
This is another part of the breakup of the massive BuildPlugin. This PR
moves the code for configuring publications to a separate plugin. Most
of the time these publications are jar files, but this also supports the
zip publication we have for integ tests.
We never do any file IO or other blocking work on the transport threads
so no tangible benefit can be derived from using more threads than CPUs
for IO.
There are however significant downsides to using more threads than necessary
with Netty in particular. Since we use the default setting for
`io.netty.allocator.useCacheForAllThreads` which is `true` we end up
using up to `16MB` of thread local buffer cache for each transport thread.
Meaning we potentially waste CPUs * 16MB of heap for unnecessary IO threads in addition to obvious inefficiencies of artificially adding extra context switches.
Two spots that allow for some optimization:
* We are often creating a composite reference of just a single item in
the transport layer => special cased via static constructor to make sure we never do that
* Also removed the pointless case of an empty composite bytes ref
* `ByteBufferReference` is practically always created from a heap buffer these days so there
is no point of dealing with all the bounds checks and extra references to sliced buffers from that
and we can just use the underlying array directly
Currently there is a clear mechanism to stub sending a request through
the transport. However, this is limited to testing exceptions on the
sender side. This commit reworks our transport related testing
infrastructure to allow stubbing request handling on the receiving side.
The use of available processors, the terminology, and the settings
around it have evolved over time. This commit cleans up some places in
the codes and in the docs to adjust to the current terminology.
This commit moves the action name validation and circuit breaking into
the InboundAggregator. This work is valuable because it lays the
groundwork for incrementally circuit breaking as data is received.
This PR includes the follow behavioral change:
Handshakes contribute to circuit breaking, but cannot be broken. They
currently do not contribute nor are they broken.
Currently all of our transport protocol decoding and aggregation occurs
in the individual transport modules. This means that each implementation
(test, netty, nio) must implement this logic. Additionally, it means
that the entire message has been read from the network before the server
package receives it.
This commit creates a pipeline in server which can be passed arbitrary
bytes to handle. Internally, the pipeline will decode, decompress, and
aggregate the messages. Additionally, this allows us to run many
megabytes of bytes through the pipeline in tests to ensure that the
logic works.
This work will enable future work:
Circuit breaking or backoff logic based on message type and byte
in the content aggregator.
Sharing bytes with the application layer using the ref counted
releasable network bytes.
Improved network monitoring based specifically on channels.
Finally, this fixes the bug where we do not circuit break on the correct
message size when compression is enabled.
Elasticsearch has a number of different BytesReference implementations.
These implementations can all implement the interface in different ways
with subtly different behavior and performance characteristics. On the
other-hand, the JVM only represents bytes as an array or a direct byte
buffer. This commit deletes the specialized Netty implementations and
moves to using a generic ByteBuffer reference type. This will allow us
to focus on standardizing performance and behave around a smaller number
of implementations that can be used by all components in Elasticsearch.
Now that the FIPS 140 security provider is simply a test dependency
we don't need the thirdPartyAudit exceptions, but plugin-cli and
transport-netty4 do need jarHell disabled as they use the non fips
BouncyCastle security provider as a test dependency too.
This change changes the way to run our test suites in
JVMs configured in FIPS 140 approved mode. It does so by:
- Configuring any given runtime Java in FIPS mode with the bundled
policy and security properties files, setting the system
properties java.security.properties and java.security.policy
with the == operator that overrides the default JVM properties
and policy.
- When runtime java is 11 and higher, using BouncyCastle FIPS
Cryptographic provider and BCJSSE in FIPS mode. These are
used as testRuntime dependencies for unit
tests and internal clusters, and copied (relevant jars)
explicitly to the lib directory for testclusters used in REST tests
- When runtime java is 8, using BouncyCastle FIPS
Cryptographic provider and SunJSSE in FIPS mode.
Running the tests in FIPS 140 approved mode doesn't require an
additional configuration either in CI workers or locally and is
controlled by specifying -Dtests.fips.enabled=true
It is the job of the http server transport to release the request in the handler
but the mock fails to do so since we never override `incomingRequest`.
* Copying the request is not necessary here. We can simply release it once the response has been generated and a lot of `Unpooled` allocations that way
* Relates #32228
* I think the issue that preventet that PR that PR from being merged was solved by #39634 that moved the bulk index marker search to ByteBuf bulk access so the composite buffer shouldn't require many additional bounds checks (I'd argue the bounds checks we add, we save when copying the composite buffer)
* I couldn't neccessarily reproduce much of a speedup from this change, but I could reproduce a very measureable reduction in GC time with e.g. Rally's PMC (4g heap node and bulk requests of size 5k saw a reduction in young GC time by ~10% for me)
Backport of #48849. Update `.editorconfig` to make the Java settings the
default for all files, and then apply a 2-space indent to all `*.gradle`
files. Then reformat all the files.
This commit introduces a consistent, and type-safe manner for handling
global build parameters through out our build logic. Primarily this
replaces the existing usages of extra properties with static accessors.
It also introduces and explicit API for initialization and mutation of
any such parameters, as well as better error handling for uninitialized
or eager access of parameter values.
Closes#42042
BytesReference is currently an abstract class which is extended by
various implementations. This makes it very difficult to use the
delegation pattern. The implication of this is that our releasable
BytesReference is a PagedBytesReference type and cannot be used as a
generic releasable bytes reference that delegates to any reference type.
This commit makes BytesReference an interface and introduces an
AbstractBytesReference for common functionality.
Elasticsearch does not grant Netty reflection access to get Unsafe. The
only mechanism that currently exists to free direct buffers in a timely
manner is to use Unsafe. This leads to the occasional scenario, under
heavy network load, that direct byte buffers can slowly build up without
being freed.
This commit disables Netty direct buffer pooling and moves to a strategy
of using a single thread-local direct buffer for interfacing with sockets.
This will reduce the memory usage from networking. Elasticsearch
currently derives very little value from direct buffer usage (TLS,
compression, Lucene, Elasticsearch handling, etc all use heap bytes). So
this seems like the correct trade-off until that changes.
Uses JDK 11's per-socket configuration of TCP keepalive (supported on Linux and Mac), see
https://bugs.openjdk.java.net/browse/JDK-8194298, and exposes these as transport settings.
By default, these options are disabled for now (i.e. fall-back to OS behavior), but we would like
to explore whether we can enable them by default, in particular to force keepalive configurations
that are better tuned for running ES.
Currently in the transport-nio work we connect and bind channels on the
a thread before the channel is registered with a selector. Additionally,
it is at this point that we set all the socket options. This commit
moves these operations onto the event-loop after the channel has been
registered with a selector. It attempts to set the socket options for a
non-server channel at registration time. If that fails, it will attempt
to set the options after the channel is connected. This should fix
#41071.
* Stop Passing Around REST Request in Multiple Spots
* Motivated by #44564
* We are currently passing the REST request object around to a large number of places. This works fine since we simply copy the full request content before we handle the rest itself which is needlessly hard on GC and heap.
* This PR removes a number of spots where the request is passed around needlessly. There are many more spots to optimize in follow-ups to this, but this one would already enable bypassing the request copying for some error paths in a follow up.
Today we have an annotation for controlling logging levels in
tests. This annotation serves two purposes, one is to control the
logging level used in tests, when such control is needed to impact and
assert the behavior of loggers in tests. The other use is when a test is
failing and additional logging is needed. This commit separates these
two concerns into separate annotations.
The primary motivation for this is that we have a history of leaving
behind the annotation for the purpose of investigating test failures
long after the test failure is resolved. The accumulation of these stale
logging annotations has led to excessive disk consumption. Having
recently cleaned this up, we would like to avoid falling into this state
again. To do this, we are adding a link to the test failure under
investigation to the annotation when used for the purpose of
investigating test failures. We will add tooling to inspect these
annotations, in the same way that we have tooling on awaits fix
annotations. This will enable us to report on the use of these
annotations, and report when stale uses of the annotation exist.
This commit moves the config that stores Cors options into the server
package. Currently both nio and netty modules must have a copy of this
config. Moving it into server allows one copy and the tests to be in a
common location.
Registering a channel with a selector is a required operation for the
channel to be handled properly. Currently, we mix the registeration with
other setup operations (ip filtering, SSL initiation, etc). However, a
fail to register is fatal. This PR modifies how registeration occurs to
immediately close the channel if it fails.
There are still two clear loopholes for how a user can interact with a
channel even if registration fails. 1. through the exception handler.
2. through the channel accepted callback. These can perhaps be improved
in the future. For now, this PR prevents writes from proceeding if the
channel is not registered.
Simplifies AbstractSimpleTransportTestCase to use JVM-local ports and also adds an assertion so
that cases like #44134 can be more easily debugged. The likely reason for that one is that a test,
which was repeated again and again while always spawning a fresh Gradle worker (due to Gradle
daemon) kept increasing Gradle worker IDs, causing an overflow at some point.
Currently nio implements ip filtering at the channel context level. This
is kind of a hack as the application logic should be implemented at the
handler level. This commit moves the ip filtering into a channel
handler. This requires adding an indicator to the channel handler to
show when a channel should be closed.
We had this as a dependency for legacy dependencies that still needed
the Log4j 1.2 API. This appears to no longer be necessary, so this
commit removes this artifact as a dependency.
To remove this dependency, we had to fix a few places where we were
accidentally relying on Log4j 1.2 instead of Log4j 2 (easy to do, since
both APIs were on the compile-time classpath).
Finally, we can remove our custom Netty logger factory. This was needed
when we were on Log4j 1.2 and handled logging in our own unique
way. When we migrated to Log4j 2 we could have dropped this
dependency. However, even then Netty would still pick up Log4j 1.2 since
it was on the classpath, thus the advantage to removing this as a
dependency now.
This is related to #27260. Currently we have a single read buffer that
is no larger than a single TLS packet. This prevents us from reading
multiple TLS packets in a single socket read call. This commit modifies
our TLS work to support reading similar to the plaintext case. The data
will be copied to a (potentially) recycled TLS packet-sized buffer for
interaction with the SSLEngine.
This fixes#41794. Currently the read timeout test queues up responses
in the netty pipeline. These responses are immediately returned in the
write call, but they are not released. This commit releases the
responses. This will cause the leak detector to quit throwing
exceptions.
This is related to #27260. Currently there is a setting
http.read_timeout that allows users to define a read timeout for the
http transport. This commit implements support for this functionality
with the transport-nio plugin. The behavior here is that a repeating
task will be scheduled for the interval defined. If there have been
no requests received since the last run and there are no inflight
requests, the channel will be closed.
This is related to #27260. Currently for the SSLDriver we allocate a
dedicated network write buffer and encrypt the data into that buffer one
buffer at a time. This requires constantly switching between encrypting
and flushing. This commit adds a dedicated outbound buffer for SSL
operations that will internally allocate new packet sized buffers as
they are need (for writing encrypted data). This allows us to totally
encrypt an operation before writing it to the network. Eventually it can
be hooked up to buffer recycling.
This commit also backports the following commit:
Handle WRAP ops during SSL read
It is possible that a WRAP operation can occur while decrypting
handshake data in TLS 1.3. The SSLDriver does not currently handle this
well as it does not have access to the outbound buffer during read call.
This commit moves the buffer into the Driver to fix this issue. Data
wrapped during a read call will be queued for writing after the read
call is complete.