Loadgen runner #12
Merged
Loadgen runner #12
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Don't sleep negative time
Capture chain storage at the end of every stage
Rename loadGen -> loadgen
Close slog fifo stream when chain process has started but swingset hasn't yet
The perf script would ignore directed signals while sleeping or waiting for docker, launch those in the background instead and wait on them. Send a term signal to docker instead of stopping to prevent abrupt kill Streamline signal handling logic
mhofman
force-pushed
the
mhofman/3107-perf-runner
branch
from
bdb6aa4 to
c0f1db6
Compare
Fix eslint-config dependency
Do not fail fast, but avoid exec races
mhofman
force-pushed
the
mhofman/3107-perf-runner
branch
from
6896aab to
94f9678
Compare
|
@warner let me know if the readme and PR description make sense. |
warner
approved these changes
Aug 17, 2021
There was a problem hiding this comment.
Yeah, that looks good.. it might be good to have an extra paragraph or two near the top as a "welcome testnet validators!" instruction guide, to distinguish between using this repo for dailyperf purposes and using it to complete the run-a-loadgen task that we've asked them to do.
|
I'm gonna fast follow up with another PR that fixes the AMM task and moves the readme, replacing it with something more appropriate. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
This (large) PR introduces a runner capable of orchestrating a chain node, a client ag-solo and the existing loadgen script, as well as several wrapper scripts to automate running the loadgen using any SDK revision, in the expected environment.
It's the base for running a daily loadgen on the latest agoric-sdk to capture historical performance data. See Agoric/agoric-sdk#2630
The orchestrator starts a local chain node (or monitor node in the case of an remote chain), pipes and parses the slog using a FIFO file, and monitors events from the slog. At regular interval, it captures process information for the CPU usage of the chain process and its vat subprocesses, as well as the disk storage used by the chain. It can run multiple cycles of chain, client and loadgen, compressing and saving the chain storage and raw slog for every stage.
Runner structure
The runner is setup as a node workspace package. Unlike the
loadgenpackage, it doesn't depend on agoric-sdk. Since it tries to follow code patterns, it runs under SES using its owninstall-sesand uses a fork ofpromise-kit, mainly to avoid breaking changes such as RESM->NESM.The runner is a CLI application composed of a
main.jsfile parsing the command line and implementing the bulk of the orchestration logic. It relies ontasksthat implement the specifics of setting up and running a chain node, running a local solo client, and running the load generator itself, all spawned as sub processes. Currently it supports running these tasks for a local chain, or against an externalstageandtestnetchain. The implementation uses a bunch of potentially reusable async, stream, process and filesystem helpers (currently not covered by any unit tests).The output of all spawned sub processes is combined and prefixed into the stdout of the runner.