bldr-daemon
Builder storage node — local block store and cluster services
Builder storage node — local block store and cluster services
Generated from bldr-daemon 0.1.0.
bldr-daemon [BLDR-DAEMON FLAGS] <COMMAND>Commands
| Command | |
|---|---|
bldr-daemon check | Validate this binary and its configuration, then exit |
bldr-daemon serve | Start the full node (networking + cluster services) and run until Ctrl-C |
Flags
--log-level <LOG_LEVEL>
Log level / filter (e.g. "info", "debug", "warn", or "mycrate=debug,hyper=warn")
Defaults to info.
--otlp-url <OTLP_URL>
OTLP collector endpoint (gRPC). Example: http://localhost:4317
--otlp-application-name <OTLP_APPLICATION_NAME>
--pyroscope-url <PYROSCOPE_URL>
--pyroscope-application-name <PYROSCOPE_APPLICATION_NAME>
--pyroscope-tag <PYROSCOPE_TAG>
May be given more than once.
--pyroscope-sample-rate <PYROSCOPE_SAMPLE_RATE>
pprof sampling frequency in Hz for the Pyroscope profiler. Higher = finer-grained profiles at more overhead. Default 100 (fine-grained, for the active performance work)
Defaults to 100.
--heap-profile
Sample the heap from startup and write a profile whenever memory crosses a new watermark. Needed to diagnose an OOM kill, which leaves nothing to ask afterwards
Defaults to false.
--heap-profile-dir <DIR>
Where profiles are written. Each is <dir>/<pid>-<pct>pct.heap
Defaults to /var/tmp/bldr-heap.
--heap-watch-interval <SECS>
How often memory is sampled, in seconds. Also the proactive-reclaim cadence: under refault churn the cgroup refills ~2 G in ~5 s (measured during the 48082 kill), so the default must not be slower than that
Defaults to 5.
--data <PATH>
Root directory for all node data (<PATH>/kv for the block store, <PATH>/keypair for the node identity)
--config <FILE>
Path to a YAML config file for networking and node settings. Missing file is silently ignored and built-in defaults are used