diff options
Diffstat (limited to 'src/tools/remote_worker/README.md')
| -rw-r--r-- | src/tools/remote_worker/README.md | 15 |
1 files changed, 10 insertions, 5 deletions
diff --git a/src/tools/remote_worker/README.md b/src/tools/remote_worker/README.md index 7664d3345a..088ea7849e 100644 --- a/src/tools/remote_worker/README.md +++ b/src/tools/remote_worker/README.md @@ -1,5 +1,6 @@ This program implements a remote execution worker that uses gRPC to accept work -requests. It also serves as a Hazelcast server for distributed caching. +requests. It can work as a remote execution worker, a cache worker, or both. +The simplest setup is as follows: - First build remote_worker and run it. @@ -9,9 +10,13 @@ requests. It also serves as a Hazelcast server for distributed caching. - Then you run Bazel pointing to the remote_worker instance. - bazel build --hazelcast_node=127.0.0.1:5701 --spawn_strategy=remote \ - --remote_worker=127.0.0.1:8080 src/tools/generate_workspace:all + bazel --host_jvm_args=-Dbazel.DigestFunction=SHA1 build \ + --spawn_strategy=remote --remote_cache=localhost:8080 \ + --remote_worker=localhost:8080 src/tools/generate_workspace:all The above command will build generate_workspace with remote spawn strategy that -uses Hazelcast as the distributed caching backend and executes work remotely on -the localhost remote_worker. +uses the local worker as the distributed caching and execution backend. + +You can also use a Hazelcast server for the distributed cache as follows: +Suppose your Hazelcast server is listening on address:port. Then, run the +remote worker with --hazelcast_node=address:port. |