@@ -9,8 +9,6 @@ many repositories) for our CI pipelines.
99 - [ Usage] ( #check_dependent_project-usage )
1010 - [ Explanation] ( #check_dependent_project-explanation )
1111 - [ Implementation] ( #check_dependent_project-implementation )
12- - [ bench-bot] ( #bench_bot )
13- - [ Infrastructure] ( #bench_bot_infra )
1412
1513# check_dependent_project <a name =" check_dependent_project " ></a >
1614
@@ -146,70 +144,3 @@ typechecking would fail.
146144
147145The check is currently limited to ` cargo check ` but optimally it should
148146[ entail the whole CI pipeline of ` $dependent_repo ` ] ( https://github.com/paritytech/ci_cd/issues/234 ) .
149-
150- # bench-bot <a name =" bench_bot " ></a >
151- The script through which command-bot runs benchmarks. See https://github.com/paritytech/command-bot#pull-request-command-queue for its usage.
152-
153- ## Infrastrcture <a name =" #bench_bot_infra " ></a >
154-
155- This bot runs on [ bm3 machine] ( https://gitlab.parity.io/parity/devops/-/blob/462e85a44e9c84064a7f91d1402e499e27e07991/inventory.yaml#L485 ) ,
156- currently maintained by CI/CD team.
157-
158- ### Configuration requirements:
159- The machine needs to configred as a production gitlab-runer.
160-
161- Config requirements:
162- * gitlab_runner_executor: shell
163- * gitlab_runner_additional_tags: bench-bot
164- * os: ubuntu20
165- * provider: ovh
166- * ` $HOME ` for ` gitlab-runner ` needs to be within ` /home ` because it is mounted on a disk with more space than the root partition;
167- the projects' release artifacts require a lot of disk space.
168- * install rustup for gitlab-runner user
169- * set default rustup
170- * install toolchain nightly
171- * add wasm target
172- * install cmake
173-
174- All those requirements are saved in an ansible role called ` rustup `
175- TODO: add link to rustup role once the PR is merged
176-
177- ### Ansible
178- Because there is only one host running code for this bot, one needs to create
179- a test VM and add it to inventory everywhere ` bm3 ` is mentioned.
180-
181- ```
182- test-bench-bot:
183- ansible_host: 34.141.84.244
184- os: ubuntu20
185- provider: ovh
186- gitlab_runner_executor: shell
187- gitlab_runner_additional_tags: bench-bot
188- ### required for rustup ansible role
189- # Install rustup, and rust toolchain as defined value.
190- rustup_user: gitlab-runner
191- # Adds {HOME/.cargo/bin} to user's PATH
192- rustup_configure_shell: true
193- # do not install any cargo crates
194- rustup_cargo_crates: []
195- ```
196-
197- ** When creating the VM helps a lot if you include your ssh public key**
198- ```
199- /home/$USER/.ssh/id_rsa_yubikey.pub
200- ```
201-
202- Test the ansible role from ` parity/devops `
203- ```
204- ansible-playbook ansible/gitlab-runner.yaml -l test-bench-bot -u radupopa2010 -t gitlab-runner-benchmarks-weights-provision
205- ```
206-
207- ### Test
208-
209- To avoid disrupting the production service while testing, a ` test-bench-bot ` configuration can be used,
210- which uses the ` test-bench-bot ` tag instead of ` bench-bot ` . It can be tested as:
211-
212- 1 . First create a dummy PR in substrate repo like [ this one] ( https://github.com/paritytech/substrate/pull/11566 )
213- 2 . Add a comment ` /cmd queue -c test-bench-bot $ pallet dev pallet_balances `
214- 3 . Wait a bit and check for a comment from the bot with the job that started
215-
0 commit comments