rls: Change AdaptiveThrottler to use Ticker instead of TimeProvider
* Use a slot being null to mark invalid rather than relying on the slot's endNanos value.
Fixes#9048
* rls: Support multiple returned targets from RLS Server
Pick the first target that is not in TRANSIENT_FAILURE state. If none, use the first target.
Also initialize all targets returned from RLS so DataCache will contain a list of child policy wrappers.
Fixes#9236
%s is fairly safe (requires a Formattable to use Locale), so %d is the
main risk item. Places that really didn't need to use String.format()
were converted to plain string concatenation. Logging locations were
generally converted to using the log infrastructure's delayed
formatting, which is generally locale-sensitive but we're okay with
that. That wasn't done in okhttp, however, because Android frequently
doesn't use MessageFormat so we'd lose the parameters. Everywhere else
was explicitly defined to be Locale.US, to be consistent independent of
the default system locale.
- The primary should've been `GKE_CLUSTER_PSM_LB`
- The secondary cluster was not activated for LB tests. This resulted
in the failover test failing, as it relies on workloads running in
different zones.
Secondary cluster was not activated for LB tests. This resulted in the failover test failing, as it relies on workloads running in different zones.
ref b/238226704
This reverts commit c1abc7f8ac. It
produced compilation issues inside Google. I strongly suspect it isn't
this commit or gRPC's fault, but it prevents further testing until it is
resolved.
Fix a bug where the server stream delivers halfClose() to the call during cancellation. It happens when call has a short deadline. Server sees `INTERNAL, desc: Half-closed without a request` due to the bug.
The two checker tasks run quickly so don't gain much from UP-TO-DATE,
but it is convenient to not see them in the noise (checkUpperBoundDeps
in particular). Gradle only performs UP-TO-DATE checks (on the inputs)
if the task has both inputs and outputs defined.
The biggest saving was for distZip/distTar/shadowDistZip/shadowDistTar
which were using the same name for the non-shadow and shadow versions.
Thus the output file would always be out-of-date because it had been
rewritten and was invalid. This is worrisome because we could have
"randomly" been using the shadow Zip/Tar at times and the non-shadow
ones at others, although I think in practice the shadow tasks always run
last and so those are the files we'd see. Changing the classifier avoids
the colliding file names. These tasks took ~7 seconds, so incremental
builds are considerably shorter now.
RuntimeOnly dependencies have been missing since 3624d59. This is
because the implementation configuration extendsFrom the shadow
configuration, so any of the things like runtimeOnly are being lost.
This change isn't "correct" but it stops the bleeding with minimal cost.
It is probably incorrect to be using shadow plugin in interop-testing at
all.
Same as #9347, but for GCE framework too (xds and xds_v3 jobs).
Should fix "Expiring Daemon because JVM heap space is exhausted".
PR #9269 probably pushed the build
over the edge, but there's been evidence via flakes for a good while
that we've been reaching the limit.
b/238334438
Android linters can't recognize the difference when VisibleForTesting
is used because the method has different visibility, or because
the method only intended for testing.
Because of that linter complains when VisibleForTesting methods are
used in the production code.
Ideally we want to replace or remove this annotation, as its
usage for marking altered visibility for testing purposes is
discouraged since guava v30.0.
INVALID_ARGUMENT is propagated to the data plane if no previous config
is available. INVALID_ARGUMENT is reserved for application use; LBs
should pretty much use UNAVAILABLE exclusively.
While most of the changes are in xds, there do not appear to be likely
xds code paths that would propagate a bad status to the data plane.
Internal policies either don't use parseLoadBalancingPolicyConfig() and
instead have their configuration objects constructed directly or are
constructed transitively through the cluster manager which uses INTERNAL
if there's a child failure. There was a worrisome hole before this
commit for StatusRuntimeExceptions received by the cluster manager, but
the audit didn't find any locations throwing such an exception.
User-selected policies produce a NACK and are protected from the
existing xds client watcher paths. The worst that appears could happen
is the channel could panic (which uses INTERNAL) if a bug let a bad
configuration through.
As defined in the gRFC [A53: Option for Ignoring xDS Resource Deletion](https://github.com/grpc/proposal/blob/master/A53-xds-ignore-resource-deletion.md).
This includes semi-related changes:
* Refactor ClientXdsClientTestBase: extract verify methods for golden resources
* Parameterize ClientXdsClientV2Test and ClientXdsClientV3Test with ignoreResourceDeletion enabled and disabled
* Add FORCE_INFO and FORCE_WARNING levels to XdsLogLevel
Should fix "Expiring Daemon because JVM heap space is exhausted".
https://github.com/grpc/grpc-java/pull/9269 probably pushed the build
over the edge, but there's been evidence via flakes for a good while
that we've been reaching the limit.
b/238334438
This can avoid creating an additional 736 tasks (previously 502 out of
1591 were not created). That's not all that important as the build time
is essentially the same, but this lets us see the poor behavior of the
protobuf plugin in our own project and increase our understanding of how
to avoid task creation when developing the plugin. Of the tasks still
being created, protobuf is the highest contributor with 165 tasks,
followed by maven-publish with 76 and appengine with 53. The remaining
59 are from our own build, but indirectly caused by maven-publish.
This fixes a regression introduced in e96d0477. The NullPointerException
only happens on client-side when some other error occurred during
handshaking.
I tried to add a test, but SerializingExecutor catches+logs the
exception and the expected behavior in the circumstance is that close()
is a noop. So the NPE was entirely benign other than annoying log
messages.
This dramatically shortens build time, even for full builds. A full
assemble of xds on my laptop goes from 1m 46s to 33s at least because
errorprone is disabled for the protos.
Previously if LDS/RDS were missing or improperly configured RPCs would
fail with "UNAVAILABLE: NameResolver returned no usable address errors".
That is very confusing and not helpful for debugging.
Ideally we'd also include the node id in this error message, but that's
a bit more involved and this is a huge improvement even without it.
b/237539851
If I didn't upgrade X there is probably a reason, but worst-case the
reason was "I was lazy." I did the easy stuff, so if upgrading caused
problems of any real sort I skipped it and moved on. The main other
reason is there's some stuff we're more conservative about upgrading,
but you can't distinguish one from the other in this commit.
* buildscripts: Fix Xmx JVM flag propagation in GRADLE_OPTS
* buildscripts: double Java memory allocation pool
To reduce periodic OOMs of the "GitHub Actions Linux Testing / tests (11) (pull_request)" job.
* istio-interop-testing: create a separate project and add istio echo server code
after removing from the grpc-interop-testing project
* add jib support
* use imported echo.proto from istio repo
* use context to propagate values from interceptor so the service's echo method has all values required to compose EchoResponse
sanjaypujare
yifeizhuang
Terry Wilson
Larry Safran
yifeizhuang
Eric Anderson
Sergii Tkachenko
Sergii Tkachenko
apolcyn
Benjamin Peterson
DNVindhya
Minsoo Cheong
Jader Alcântara