This makes ClusterManagerLB more straight-forward, focusing on just the
things that are relevant to it, and it avoids specialized map key
handling in updateChildrenWithResolvedAddresses().
The child policy config should be refreshed every address update, so it
shouldn't be stored in the ChildLbState. In addition, none of the
current usages actually used what was stored in the ChildLbState in a
meaningful way (it was always null).
ResolvedAddresses was also removed from createChildLbState(), as nothing
in it should be needed for creation; it varies over time and the values
passed at creation are immutable.
Bazel had the dependency added because of #5046, where Guava was
depending on it as compile-only and Bazel build have "unknown enum
constant" warnings. Guava now has a compile dependency on j2objc, so
this workaround is no longer needed. There are currently no version skew
issues in Gradle, which was the only usage.
Since 04474970 RingHashLB has not used
acceptResolvedAddressesInternal(). At the time that was needed because
deactivated children were part of MultiChildLB. But in 9de8e443, the
logic of RingHashLB and MultiChildLB.acceptResolvedAddressesInternal()
converged, so it can now swap back to using the base class for more
logic.
One LB no longer needs to extend ChildLbState and one has to start, so
it is a bit of a wash. There are more LBs that need the auto-request
logic, but if we have an API where subclasses override it without
calling super then we can't change the implementation in the future.
Adding behavior on top of a base class allows subclasses to call super,
which lets the base class change over time.
Some addresses are equal even though their toString is different
(InetSocketAddress ignores the hostname when it has an address). And
some addresses are not equal even though their toString might be the
same (AnonymousInProcessSocketAddress doesn't override toString()).
InetSocketAddress/InetAddress do not cache the toString() result. Thus,
even in the worst case that uses a HashSet, this should use less memory
than the earlier approach, as no strings are formatted. It probably also
significantly improves performance in the reasonably common case when an
Endpoint is created just for looking up a key, because the string
creation in the constructor isn't then amorized.
updateChildrenWithResolvedAddresses(), for example, creates n^2 Endpoint
objects for lookups.
They share very little code, and we really don't want RoundRobinLb to be
public and non-final. Originally, WRR was expected to share much more
code with RR, and even delegated to RR at times. The delegation was
removed in 111ff60e. After dca89b25, most of the sharing has been moved
out into general-purpose tools that can be used by any LB policy.
FixedResultPicker now has equals to makes it as a EmptyPicker
replacement. RoundRobinLb still uses EmptyPicker because fixing its
tests is a larger change. OutlierDetectionLbTest was changed because
FixedResultPicker is used by PickFirstLeafLb, and now RoundRobinLb can
squelch some of its updates for ready pickers.
This is to replace switchTo(), to allow composing GracefulSwitchLb with
other helpers like MultiChildLb. It also prevents users of
GracefulSwitchLb from needing to use ServiceConfigUtil.
This will be used by the metadata exchange of CSM. When recording
per-attempt metrics, we really need per-attempt data and can't leverage
ClientInterceptors.
gRFC A78 has WRR and pick-first include a `grpc.target` label, defined
in A66:
> `grpc.target` : Canonicalized target URI used when creating gRPC
> Channel, e.g. "dns:///pubsub.googleapis.com:443",
> "xds:///helloworld-gke:8000". Canonicalized target URI is the form
> with the scheme included if the user didn't mention the scheme
> (`scheme://[authority]/path`). For channels such as inprocess channels
> where a target URI is not available, implementations can synthesize a
> target URI.
As part of gRFC A78:
> To support the locality label in the per-call metrics, we will provide
> a mechanism for LB picker to add optional labels to the call attempt
> tracer.
It is easy to manage these things outside of MultiChildLb and it makes
the shared code easier and use less memory. In particular, we don't want
to use many instances of GracefulSwitchLb in virtually every policy
simply because it was needed in one or two cases.
Adds interfaces required for recording metrics from gRPC components. And added API to get `MetricRecorder` in `LoadBalancer.Helper` and add `MetricSink` to `ManagedChannelBuilder`.
The recommended way to load dependencies from `rules_jvm_external`
is to make use of the `@maven` workspace, and the most readable
way of doing that is to use the `artifact` macro provides.
This removes the need to generate the "compat" namespaces, which
`rules_jvm_external` provided for backwards compatibility with
older releases. This change also sets things up for supporting
`bzlmod`: this requires all workspaces accessed by a library to
be named "up front" in the `MODULE.bazel` file. This way, the
only repo that needs to be exported is `@maven`, rather than the
current huge list.
Including a Status description makes it easier to debug subchannel
closure issues if it's clear that a subchannel became unavailable because
of an outlier detection ejection.
* Allow the queued byte threshold for a Stream to be ready to be configurable
- on clients this is exposed by setting a CallOption
- on servers this is configured by calling a method on ServerCall or ServerStreamListener
* add final, change method permissions, add javadoc, cleanup unneeded, move updateOverallBalancingState to ClusterManagerLB and make it abstract
* Restructure to eliminate the flags as protected methods
* Move methods around so that the candidates for override are near the top.
* Reorder picker methods lower
This change has health checking consumer (new pick first) to install a listener through and health checking producer (outlier detection and client health checking) producing health checks. Health notification chain is built reusing the previous connectivity state chain.
Pickfirst installs the health listener, and is capable of detecting when no health checking producer is installed in the system. In that case, it sets health status to be READY so that health system is no-op.
We already do this for WRR. Notably, we are no longer trying to avoid
the modulus each pick. It was of questionable value, and removing it is
necessary to continue sharing the same integer when the list size
changes.
The change means we can implement a stronger isEquivalentTo() by
comparing the AtomicInteger references. It is strong enough that the
operation aligns with normal equals(). Using equals() instead of
isEquivalentTo() also made more obvious an equals() optimization that
uses the hashCode() before the more expensive HashSet creation; equals()
should now be very fast except when they are (very likely) equal.
Vindhya Ningegowda
Eric Anderson
Kurt Alfred Kluever
erm-g
cfredri4
Andrey Ermolov
Matthew Stevenson
Daniel Liu
David Burns
Terry Wilson
Larry Safran
James Duong
yifeizhuang
Terry Wilson