build: override grpc dependencies with our dependencies

Don't get transitive grpc dependencies, use the ones from our `maven_install(...)` chore(deps): bump protobuf runtime to 3.19.1 chore(deps) add transitive dependencies feat: add Proto reflection service to shard worker To aid connection troubleshooting Bug: Fix Blocked thread in WriteStreamObserver Caused by CASFile Write (bazelbuild#1486) * Add unit test * Signal Write on complete Pin the Java toolchain to `remotejdk_17` (bazelbuild#1509) Closes bazelbuild#1508 Cleanups: - remove the unused `ubuntu-bionic` base image - replace `ubuntu-jammy:jammy-java11-gcc` with `ubuntu-mantic:mantic-java17-gcc` - replace `amazoncorretto:19` with `ubuntu-mantic:mantic-java17-gcc` - swap inverted log file names in a file docs: add markdown language specifiers for code blocks Support OutputPaths in OutputDirectory Specifying any number of OutputPaths will ignore OutputFiles (consistent with uploads). Where an OutputPath specifies an output directory, the action must be able to create the directory itself. Permit Absolute Symlink Targets with configuration Partial specification of the absolute symlink response per REAPI. Remaining work will be in output identification. chore: update bazel to 6.4.0 (bazelbuild#1513) Trying to get more info on the Lombok stamping issue on Windows CI. See also bazelbuild/bazel#10363 and bazelbuild/bazel#18185 Rename instance types (bazelbuild#1514) Create SymlinkNode outputs during upload (bazelbuild#1515) Default disabled, available with createSymlinkOutputs option in Worker config. feat: Implement CAS lease extension (bazelbuild#1455) Problem Enabling the findMissingBlobsViaBackplane flag in BuildfarmServer eliminates the need for the BuildfarmWorker's fmb API call. This BuildfarmWorker:fmb call was also responsible for tracking CAS entry access. As result, our CAS cache eviction strategy shifted from LRU to FIFO. When the findMissingBlobsViaBackplane flag is enabled, the buildfarm relies on the backplane as the definitive source for CAS availability. Since we don't update CAS expiry on each access, the backplane will independently expire CAS entries based on the specified cas_expire duration, even if they are actively being read. Solution Updated bfServer:fmb call to perform non-blocking fmb calls to workers, allowing these workers to record access for the relevant CAS entries. Extended expiry duration for available CAS entries in the backplane on each fmb call. With these changes, we can utilize Bazel's experimental_remote_cache_lease_extension and experimental_remote_cache_ttl flags for incremental builds. Closes bazelbuild#1428 Bump org.json:json from 20230227 to 20231013 in /admin/main (bazelbuild#1516) Bumps [org.json:json](https://github.com/douglascrockford/JSON-java) from 20230227 to 20231013. - [Release notes](https://github.com/douglascrockford/JSON-java/releases) - [Changelog](https://github.com/stleary/JSON-java/blob/master/docs/RELEASES.md) - [Commits](https://github.com/douglascrockford/JSON-java/commits) --- updated-dependencies: - dependency-name: org.json:json dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Re-add missing graceful shutdown functionality (bazelbuild#1520) Technically correct to unwrap EE on lock failure Bump rules_oss_audit and patch for py3.11 Prevent healthStatusManager NPE on start failure Consistent check for publicName presence Read through external with query THROUGH=true Specifying a correlated invocation id with a uri containing a THROUGH=true query param will cause the CFC to read a blob through an external input stream, populating locally along the way. This permits client-based replication of blobs, and can enable N+1 replication and traffic balancing for reads. Add --port option to worker Option to run the worker with a cmdline specification for its gRPC server port. Restore worker --root cmdline specification Root cmdline specification has been broken since the config change of v2. Make bf-executor small blob names consistent Remove the size identification for small blobs when uploading with bf-executor. feat: Hot CAS Entries - Update read counts in Redis
amishra-u · Nov 8, 2023 · 8ca62bd · 8ca62bd
1 parent 1244187
commit 8ca62bd
Show file tree

Hide file tree

Showing 51 changed files with 777 additions and 213 deletions.
diff --git a/.bazelci/run_server_test.sh b/.bazelci/run_server_test.sh
@@ -8,11 +8,11 @@ bazel build //src/main/java/build/buildfarm:buildfarm-shard-worker
 bazel build //src/main/java/build/buildfarm:buildfarm-server
 
 # Start a single worker
-bazel run //src/main/java/build/buildfarm:buildfarm-shard-worker $(pwd)/examples/config.minimal.yml > server.log 2>&1 &
+bazel run //src/main/java/build/buildfarm:buildfarm-shard-worker $(pwd)/examples/config.minimal.yml > worker.log 2>&1 &
 echo "Started buildfarm-shard-worker..."
 
 # Start a single server
-bazel run //src/main/java/build/buildfarm:buildfarm-server $(pwd)/examples/config.minimal.yml > worker.log 2>&1 &
+bazel run //src/main/java/build/buildfarm:buildfarm-server $(pwd)/examples/config.minimal.yml > server.log 2>&1 &
 echo "Started buildfarm-server..."
 
 echo "Wait for startup to finish..."

diff --git a/.bazelrc b/.bazelrc
@@ -1,3 +1,9 @@
+build --java_language_version=17
+build --java_runtime_version=remotejdk_17
+
+build --tool_java_language_version=17
+build --tool_java_runtime_version=remotejdk_17
+
 common --enable_platform_specific_config
 
 build:fuse --define=fuse=true

diff --git a/.bazelversion b/.bazelversion
@@ -1 +1 @@
-6.2.0
+6.4.0
diff --git a/BUILD b/BUILD
@@ -120,7 +120,7 @@ sh_binary(
 java_image(
     name = "buildfarm-server",
     args = ["/app/build_buildfarm/examples/config.minimal.yml"],
-    base = "@amazon_corretto_java_image_base//image",
+    base = "@ubuntu-mantic//image",
     classpath_resources = [
         "//src/main/java/build/buildfarm:configs",
     ],
@@ -148,14 +148,14 @@ oss_audit(
 # Download cgroup-tools so that the worker is able to restrict actions via control groups.
 download_pkgs(
     name = "worker_pkgs",
-    image_tar = "@ubuntu-jammy//image",
+    image_tar = "@ubuntu-mantic//image",
     packages = ["cgroup-tools"],
     tags = ["container"],
 )
 
 install_pkgs(
     name = "worker_pkgs_image",
-    image_tar = "@ubuntu-jammy//image",
+    image_tar = "@ubuntu-mantic//image",
     installables_tar = ":worker_pkgs.tar",
     installation_cleanup_commands = "rm -rf /var/lib/apt/lists/*",
     output_image_name = "worker_pkgs_image",

diff --git a/README.md b/README.md
@@ -19,17 +19,17 @@ All commandline options override corresponding config settings.
 
 Run via
 
-```
-docker run -d --rm --name buildfarm-redis -p 6379:6379 redis:5.0.9
+```shell
+$ docker run -d --rm --name buildfarm-redis -p 6379:6379 redis:5.0.9
 redis-cli config set stop-writes-on-bgsave-error no
 ```
 
 ### Bazel Buildfarm Server
 
 Run via
 
-```
-bazelisk run //src/main/java/build/buildfarm:buildfarm-server -- <logfile> <configfile>
+```shell
+$ bazelisk run //src/main/java/build/buildfarm:buildfarm-server -- <logfile> <configfile>
 
 Ex: bazelisk run //src/main/java/build/buildfarm:buildfarm-server -- --jvm_flag=-Djava.util.logging.config.file=$PWD/examples/logging.properties $PWD/examples/config.minimal.yml
 ```
@@ -40,8 +40,8 @@ Ex: bazelisk run //src/main/java/build/buildfarm:buildfarm-server -- --jvm_flag=
 
 Run via
 
-```
-bazelisk run //src/main/java/build/buildfarm:buildfarm-shard-worker -- <logfile> <configfile>
+```shell
+$ bazelisk run //src/main/java/build/buildfarm:buildfarm-shard-worker -- <logfile> <configfile>
 
 Ex: bazelisk run //src/main/java/build/buildfarm:buildfarm-shard-worker -- --jvm_flag=-Djava.util.logging.config.file=$PWD/examples/logging.properties $PWD/examples/config.minimal.yml
 
@@ -53,9 +53,9 @@ Ex: bazelisk run //src/main/java/build/buildfarm:buildfarm-shard-worker -- --jvm
 
 To use the example configured buildfarm with bazel (version 1.0 or higher), you can configure your `.bazelrc` as follows:
 
-```
+```shell
 $ cat .bazelrc
-build --remote_executor=grpc://localhost:8980
+$ build --remote_executor=grpc://localhost:8980
 ```
 
 Then run your build as you would normally do.
@@ -67,20 +67,20 @@ Buildfarm uses [Java's Logging framework](https://docs.oracle.com/javase/10/core
 You can use typical Java logging configuration to filter these results and observe the flow of executions through your running services.
 An example `logging.properties` file has been provided at [examples/logging.properties](examples/logging.properties) for use as follows:
 
-```
-bazel run //src/main/java/build/buildfarm:buildfarm-server -- --jvm_flag=-Djava.util.logging.config.file=$PWD/examples/logging.properties $PWD/examples/config.minimal.yml
+```shell
+$ bazel run //src/main/java/build/buildfarm:buildfarm-server -- --jvm_flag=-Djava.util.logging.config.file=$PWD/examples/logging.properties $PWD/examples/config.minimal.yml
 ```
 
 and
 
-```
-bazel run //src/main/java/build/buildfarm:buildfarm-shard-worker -- --jvm_flag=-Djava.util.logging.config.file=$PWD/examples/logging.properties $PWD/examples/config.minimal.yml
+``` shell
+$ bazel run //src/main/java/build/buildfarm:buildfarm-shard-worker -- --jvm_flag=-Djava.util.logging.config.file=$PWD/examples/logging.properties $PWD/examples/config.minimal.yml
 ```
 
 To attach a remote debugger, run the executable with the `--debug=<PORT>` flag. For example:
 
-```
-bazel run //src/main/java/build/buildfarm:buildfarm-server -- --debug=5005 $PWD/examples/config.minimal.yml
+```shell
+$ bazel run //src/main/java/build/buildfarm:buildfarm-server -- --debug=5005 $PWD/examples/config.minimal.yml
 ```
 
 

diff --git a/_site/docs/configuration/configuration.md b/_site/docs/configuration/configuration.md
@@ -7,7 +7,7 @@ has_children: true
 
 Minimal required:
 
-```
+```yaml
 backplane:
   redisUri: "redis://localhost:6379"
   queues:
@@ -28,17 +28,18 @@ For an example configuration containing all of the configuration values, see `ex
 
 ### Common
 
-| Configuration        | Accepted and _Default_ Values | Command Line Argument | Description                                       |
-|----------------------|-------------------------------|-----------------------|---------------------------------------------------|
-| digestFunction       | _SHA256_, SHA1                |                       | Digest function for this implementation           |
-| defaultActionTimeout | Integer, _600_                |                       | Default timeout value for an action (seconds)     |
-| maximumActionTimeout | Integer, _3600_               |                       | Maximum allowed action timeout (seconds)          |
-| maxEntrySizeBytes    | Long, _2147483648_            |                       | Maximum size of a single blob accepted (bytes)    |
-| prometheusPort       | Integer, _9090_               | --prometheus_port     | Listening port of the Prometheus metrics endpoint |
+| Configuration                | Accepted and _Default_ Values | Command Line Argument | Description                                                  |
+|------------------------------|-------------------------------|-----------------------|--------------------------------------------------------------|
+| digestFunction               | _SHA256_, SHA1                |                       | Digest function for this implementation                      |
+| defaultActionTimeout         | Integer, _600_                |                       | Default timeout value for an action (seconds)                |
+| maximumActionTimeout         | Integer, _3600_               |                       | Maximum allowed action timeout (seconds)                     |
+| maxEntrySizeBytes            | Long, _2147483648_            |                       | Maximum size of a single blob accepted (bytes)               |
+| prometheusPort               | Integer, _9090_               | --prometheus_port     | Listening port of the Prometheus metrics endpoint            |
+| allowSymlinkTargetAbsolute   | boolean, _false_              |                       | Permit inputs to contain symlinks with absolute path targets |
 
 Example:
 
-```
+```yaml
 digestFunction: SHA1
 defaultActionTimeout: 1800
 maximumActionTimeout: 1800
@@ -79,7 +80,7 @@ worker:
 
 Example:
 
-```
+```yaml
 server:
   instanceType: SHARD
   name: shard
@@ -96,7 +97,7 @@ server:
 
 Example:
 
-```
+```yaml
 server:
   grpcMetrics:
     enabled: false
@@ -114,7 +115,7 @@ server:
 
 Example:
 
-```
+```yaml
 server:
   caches:
     directoryCacheMaxEntries: 10000
@@ -132,7 +133,7 @@ server:
 
 Example:
 
-```
+```yaml
 server:
   admin:
     deploymentEnvironment: AWS
@@ -151,14 +152,14 @@ server:
 
 Example:
 
-```
+```yaml
 server:
   metrics:
     publisher: log
     logLevel: INFO
 ```
 
-```
+```yaml
 server:
   metrics:
     publisher: aws
@@ -207,7 +208,7 @@ server:
 
 Example:
 
-```
+```yaml
 backplane:
   type: SHARD
   redisUri: "redis://localhost:6379"
@@ -224,7 +225,7 @@ backplane:
 
 Example:
 
-```
+```yaml
 backplane:
   type: SHARD
   redisUri: "redis://localhost:6379"
@@ -261,8 +262,9 @@ backplane:
 | errorOperationRemainingResources | boolean, _false_              |                       |                                                                                                                                                                                                                                                                                                                          |
 | realInputDirectories             | List of Strings, _external_   |                       | A list of paths that will not be subject to the effects of linkInputDirectories setting, may also be used to provide writable directories as input roots for actions which expect to be able to write to an input location and will fail if they cannot                                                                  |
 | gracefulShutdownSeconds          | Integer, 0                    |                       | Time in seconds to allow for operations in flight to finish when shutdown signal is received                                                                                                                                                                                                                             |
+| createSymlinkOutputs             | boolean, _false_              |                       | Creates SymlinkNodes for symbolic links discovered in output paths for actions. No verification of the symlink target path occurs. Buildstream, for example, requires this.                                                                                                                                              |
 
-```
+```yaml
 worker:
   port: 8981
   publicName: "localhost:8981"
@@ -279,7 +281,7 @@ worker:
 
 Example:
 
-```
+```yaml
 worker:
   capabilities:
     cas: true
@@ -296,7 +298,7 @@ worker:
 
 Example:
 
-```
+```yaml
 worker:
   sandboxSettings:
     alwaysUse: true
@@ -313,7 +315,7 @@ worker:
 
 Example:
 
-```
+```yaml
 worker:
   dequeueMatchSettings:
     acceptEverything: true
@@ -333,7 +335,7 @@ worker:
 
 Example:
 
-```
+```yaml
 worker:
   storages:
     - type: FILESYSTEM
@@ -361,7 +363,7 @@ worker:
 
 Example:
 
-```
+```yaml
 worker:
   executionPolicies:
     - name: test

diff --git a/_site/docs/execution/execution_policies.md b/_site/docs/execution/execution_policies.md
@@ -17,7 +17,7 @@ This policy type specifies that a worker should prepend a single path, and a num
 
 This example will use the buildfarm-provided executable `as-nobody`, which will upon execution demote itself to a `nobody` effective process owner uid, and perform an `execvp(2)` with the remaining provided program arguments, which will subsequently execute as a user that no longer matches the worker process.
 
-```
+```yaml
 # default wrapper policy application
 worker:
   executionPolicies:
@@ -50,32 +50,37 @@ These wrappers are used for detecting actions that rely on time.  Below is a dem
 This addresses two problems in regards to an action's dependence on time.  The 1st problem is when an action takes longer than it should because it's sleeping unnecessarily.  The 2nd problem is when an action relies on time which causes it to eventually be broken on master despite the code not changing.  Both problems are expressed below as unit tests.  We demonstrate a time-spoofing mechanism (the re-writing of syscalls) which allows us to detect these problems generically over any action.  The objective is to analyze builds for performance inefficiency and discover future instabilities before they occur.
 
 ### Issue 1 (slow test)
-```
+
+```bash
 #!/bin/bash
 set -euo pipefail
 
 echo -n "testing... "
 sleep 10;
 echo "done"
 ```
+
 The test takes 10 seconds to run on average.
-```
-bazel test --runs_per_test=10 --config=remote //cloud/buildfarm:sleep_test
+
+```shell
+$ bazel test --runs_per_test=10 --config=remote //cloud/buildfarm:sleep_test
 //cloud/buildfarm:sleep_test                                             PASSED in 10.2s
   Stats over 10 runs: max = 10.2s, min = 10.1s, avg = 10.2s, dev = 0.0s
 ```
 
 We can check for performance improvements by using the `skip-sleep` option.
-```
-bazel test --runs_per_test=10 --config=remote --remote_default_exec_properties='skip-sleep=true' //cloud/buildfarm:sleep_test
+
+```shell
+$ bazel test --runs_per_test=10 --config=remote --remote_default_exec_properties='skip-sleep=true' //cloud/buildfarm:sleep_test
 //cloud/buildfarm:sleep_test                                             PASSED in 1.0s
   Stats over 10 runs: max = 1.0s, min = 0.9s, avg = 1.0s, dev = 0.0s
 ```
 
 Now the test is 10x faster.  If skipping sleep makes an action perform significantly faster without affecting its success rate, that would warrant further investigation into the action's implementation.
 
 ### Issue 2 (future failing test)
-```
+
+```bash
 #!/bin/bash
 set -euo pipefail
 
@@ -89,12 +94,15 @@ echo "Times change."
 date
 exit -1;
 ```
+
 The test passes today, but will it pass tomorrow?  Will it pass a year from now?  We can find out by using the `time-shift` option.
-```
-bazel test --test_output=streamed --remote_default_exec_properties='time-shift=31556952' --config=remote //cloud/buildfarm:future_fail
+
+```shell
+$ bazel test --test_output=streamed --remote_default_exec_properties='time-shift=31556952' --config=remote //cloud/buildfarm:future_fail
 INFO: Found 1 test target...
 Times change.
 Mon Sep 25 18:31:09 UTC 2023
 //cloud/buildfarm:future_fail                                            FAILED in 18.0s
 ```
+
 Time is shifted to the year 2023 and the test now fails.  We can fix the problem before others see it.