This change has these main aspects to it:
1. Removal of any name resolution responsibility from ManagedChannelImpl
2. Creation of a new RetryScheduler to own generic retry logic
- Can also be used outside the name resolution context
3. Creation of a new RetryingNameScheduler that can be used to wrap any
polling name resolver to add retry capability
4. A new facility in NameResolver to allow implementations to notify
listeners on the success of name resolution attempts
- RetryingNameScheduler relies on this
* xds:Change timer creation logic to wait until the adsStream is ready before creating the timer to mark resources absent.
* xds:When the ads stream is closed only send errors to subscribers that haven't yet gotten results to match spec.
* Use a blocking queue to avoid the 2-second sleep.
For some inexplicable reason the following call.verifyRequest fails only for the V2 test and only from command line not IDE unless there is some Thread.sleep, even if it is only 1-millis.
Replace ctx.host_configuration.host_path_separator with ctx.configuration.host_path_separator.
This is because
1. we're removing ctx.host_configuration
2. host_path_separator's value doesn't depend on the configuration (the API is misleading by embedding it under ctx.configuration). So this is a no-op.
Internal ref cl/494008561.
Fix ConcurrentModificationException in PriorityLoadBalancer by making copy of children values to iterate rather than directly using children in for loop.
We use state-of-the-world approach. For LDS/CDS, the control plane must return all resources that the client has subscribed to in each request. If some LDS/CDS resources are gone in a new update, their corresponding RDS/EDS resources names will be onAbsent(), unless there is cached data that is in use by other subscribers in other components.
The motivations to remove this "retained resource" logic between resource types are:
1. Already handled by the subscribers, e.g. a CDS state would shut down its childLBs on new updates. XdsResolver for LdsUpdate would cancel all existing RDS subscriptions. Therefore the onAbsent() notification is effectively no-op.
2. Complexity.
updateAndGet is only available in API level 24+. It is unclear why
AnimalSniffer didn't detect this.
This was noticed when doing the import, but the Android CI has also been
failing because of it.
ClusterImplLoadBalancer adds the ATTR_CLUSTER_NAME and
ATTR_SSL_CONTEXT_PROVIDER_SUPPLIER attributes to the EAG list when it
creates a new subchannel, but they are lost on subsequent address
updates. This change assures the attributes are also included on address
updates.
Add big negative integer to pending stream count when cancelled. The count is used to delay closing master listener until streams fully drained.
Increment pending stream count before creating one. The count is also used to indicate callExecutor is safe to be used. New stream will not be created if big negative number was added, i.e. stream cancelled. New stream is created if not cancelled, callExecutor is safe to be used, because cancel will be delayed.
Create new streams (retry, hedging) is moved to the main thread, before callExecutor calls drain.
Minor refactor the masterListener.close() scenario.
* Report the values that were compared rather than the underlying durations that generated those values.
* Eliminate BigInteger to eliminate flakiness from possible class loading overhead.
This makes it more obvious when plaintext vs TLS is being used and is
the preferred API.
I did not change the Google Auth example, because it is doing things a
weird way and changing it would be more invasive. I also didn't update
the Android examples.
Although CallOptions is annotated by @Immutable, its fields are not
final. So it's not truly immutable, namely not safe for unsynchronized
publication.
This commit adds final to all fields of CallOptions. Using internal
builder class to keep flexibility of constructing CallOptions.
Fixes#9658
This is a potential optimization, depending on the OpenCensus
implementation installed. Currently, the code checks TagContext equality
against the empty TagContext as an optimization to avoid updating the
Context, but checking equality may not be cheaper than updating the
Context. In particular, this condition is almost always true, because we
[update the parentCtx with an additional
tag](bacf18db8d/census/src/main/java/io/grpc/census/CensusStatsModule.java (L770-L774)).
It's only true when there is no OpenCensus implementation (i.e. some
kind of no-op implementation that ignores TagContexts) or when the empty
TagContext already has that tag set to the specific value which varies
depending on the RPC method.
For the default OpenCensus implementation, the equality check is
equality between two maps. Other implementations may require additional
work to determine equality.
Basically, we are trading off rarely avoiding updating the Context for
always doing at least a map equality or potentially even more depending
on the OpenCensus implementation in use. Given this, it makes sense to
just always update the Context.
This change migrates from `cfg = "host"` to `cfg = "exec"`.
This is a no-op cleanup: `cfg = "host"` is a deprecated alias
for `cfg = "exec"`.
Tested:
```
bazel build //grpclb:grpclb
bazel build //compiler:java_grpc_library_toolchain
bazel build //compiler:grpc_java_plugin
bazel build '//compiler:*'
```
Internal ref cl/487924699
I would have assumed a JMH plugin update renamed include to includes,
but I don't see a mention to it in the plugin changelog. Either it was
an undocumented rename, or devs just kept fixing it locally and not
fixing upstream. In either case, it works now.
If a child policy triggers an update to the parent priority policy
it will be ignored if an update is already in process.
This is the second attempt to make this change, the first one caused a
problem with the ring hash LB. A new test that uses actual control plane
and data plane servers is now included to prove the issue no longer
appears.
This extracts the startup and shutdown code for the control and data
plane server to reparate JUnit rules, which allows this logic to be
resued in other tests in a simple manner. Also makes the test easier to
read with the boiler plate init code removed.
Larry Safran
Eric Anderson
Kun Zhang
Terry Wilson
Greg
yifeizhuang
Mohan Li
Smilencer
cbianchi-7
cbianchi-7
apolcyn
markb74
Sergii Tkachenko
pandaapo
Justin Bassett
RapperCL
RapperCL
yifeizhuang