#!/usr/bin/env python3 import time import logging import os import sys import boto3 from github import Github import requests from env_helper import REPO_COPY, TEMP_PATH from stopwatch import Stopwatch from upload_result_helper import upload_results from s3_helper import S3Helper from get_robot_token import get_best_robot_token, get_parameter_from_ssm from pr_info import PRInfo from compress_files import compress_fast from commit_status_helper import post_commit_status from clickhouse_helper import ClickHouseHelper, prepare_tests_results_for_clickhouse from version_helper import get_version_from_repo from tee_popen import TeePopen from ssh import SSHKey from build_download_helper import get_build_name_for_check from rerun_helper import RerunHelper JEPSEN_GROUP_NAME = "jepsen_group" DESIRED_INSTANCE_COUNT = 3 IMAGE_NAME = "clickhouse/keeper-jepsen-test" CHECK_NAME = "ClickHouse Keeper Jepsen" SUCCESSFUL_TESTS_ANCHOR = "# Successful tests" INTERMINATE_TESTS_ANCHOR = "# Indeterminate tests" CRASHED_TESTS_ANCHOR = "# Crashed tests" FAILED_TESTS_ANCHOR = "# Failed tests" def _parse_jepsen_output(path): test_results = [] current_type = "" with open(path, "r") as f: for line in f: if SUCCESSFUL_TESTS_ANCHOR in line: current_type = "OK" elif INTERMINATE_TESTS_ANCHOR in line or CRASHED_TESTS_ANCHOR in line: current_type = "ERROR" elif FAILED_TESTS_ANCHOR in line: current_type = "FAIL" if ( line.startswith("store/clickhouse-keeper") or line.startswith("clickhouse-keeper") ) and current_type: test_results.append((line.strip(), current_type)) return test_results def get_autoscaling_group_instances_ids(asg_client, group_name): group_description = asg_client.describe_auto_scaling_groups( AutoScalingGroupNames=[group_name] ) our_group = group_description["AutoScalingGroups"][0] instance_ids = [] for instance in our_group["Instances"]: if ( instance["LifecycleState"] == "InService" and instance["HealthStatus"] == "Healthy" ): instance_ids.append(instance["InstanceId"]) return instance_ids def get_instances_addresses(ec2_client, instance_ids): ec2_response = ec2_client.describe_instances(InstanceIds=instance_ids) instance_ips = [] for instances in ec2_response["Reservations"]: for ip in instances["Instances"]: instance_ips.append(ip["PrivateIpAddress"]) return instance_ips def prepare_autoscaling_group_and_get_hostnames(): asg_client = boto3.client("autoscaling", region_name="us-east-1") asg_client.set_desired_capacity( AutoScalingGroupName=JEPSEN_GROUP_NAME, DesiredCapacity=DESIRED_INSTANCE_COUNT ) instances = get_autoscaling_group_instances_ids(asg_client, JEPSEN_GROUP_NAME) counter = 0 while len(instances) < DESIRED_INSTANCE_COUNT: time.sleep(5) instances = get_autoscaling_group_instances_ids(asg_client, JEPSEN_GROUP_NAME) counter += 1 if counter > 30: raise Exception("Cannot wait autoscaling group") ec2_client = boto3.client("ec2", region_name="us-east-1") return get_instances_addresses(ec2_client, instances) def clear_autoscaling_group(): asg_client = boto3.client("autoscaling", region_name="us-east-1") asg_client.set_desired_capacity( AutoScalingGroupName=JEPSEN_GROUP_NAME, DesiredCapacity=0 ) instances = get_autoscaling_group_instances_ids(asg_client, JEPSEN_GROUP_NAME) counter = 0 while len(instances) > 0: time.sleep(5) instances = get_autoscaling_group_instances_ids(asg_client, JEPSEN_GROUP_NAME) counter += 1 if counter > 30: raise Exception("Cannot wait autoscaling group") def save_nodes_to_file(instances, temp_path): nodes_path = os.path.join(temp_path, "nodes.txt") with open(nodes_path, "w") as f: f.write("\n".join(instances)) f.flush() return nodes_path def get_run_command( ssh_auth_sock, ssh_sock_dir, pr_info, nodes_path, repo_path, build_url, result_path, docker_image, ): return ( f"docker run --network=host -v '{ssh_sock_dir}:{ssh_sock_dir}' -e SSH_AUTH_SOCK={ssh_auth_sock} " f"-e PR_TO_TEST={pr_info.number} -e SHA_TO_TEST={pr_info.sha} -v '{nodes_path}:/nodes.txt' -v {result_path}:/test_output " f"-e 'CLICKHOUSE_PACKAGE={build_url}' -v '{repo_path}:/ch' -e 'CLICKHOUSE_REPO_PATH=/ch' -e NODES_USERNAME=ubuntu {docker_image}" ) if __name__ == "__main__": logging.basicConfig(level=logging.INFO) stopwatch = Stopwatch() pr_info = PRInfo() logging.info( "Start at PR number %s, commit sha %s labels %s", pr_info.number, pr_info.sha, pr_info.labels, ) if pr_info.number != 0 and "jepsen-test" not in pr_info.labels: logging.info("Not jepsen test label in labels list, skipping") sys.exit(0) gh = Github(get_best_robot_token()) rerun_helper = RerunHelper(gh, pr_info, CHECK_NAME) if rerun_helper.is_already_finished_by_status(): logging.info("Check is already finished according to github status, exiting") sys.exit(0) if not os.path.exists(TEMP_PATH): os.makedirs(TEMP_PATH) result_path = os.path.join(TEMP_PATH, "result_path") if not os.path.exists(result_path): os.makedirs(result_path) instances = prepare_autoscaling_group_and_get_hostnames() nodes_path = save_nodes_to_file(instances, TEMP_PATH) # always use latest docker_image = IMAGE_NAME build_name = get_build_name_for_check(CHECK_NAME) if pr_info.number == 0: version = get_version_from_repo() release_or_pr = f"{version.major}.{version.minor}" else: # PR number for anything else release_or_pr = str(pr_info.number) # This check run separately from other checks because it requires exclusive # run (see .github/workflows/jepsen.yml) So we cannot add explicit # dependency on a build job and using busy loop on it's results. For the # same reason we are using latest docker image. build_url = f"https://s3.amazonaws.com/clickhouse-builds/{release_or_pr}/{pr_info.sha}/{build_name}/clickhouse" head = requests.head(build_url) counter = 0 while head.status_code != 200: time.sleep(10) head = requests.head(build_url) counter += 1 if counter >= 180: logging.warning("Cannot fetch build in 30 minutes, exiting") sys.exit(0) with SSHKey(key_value=get_parameter_from_ssm("jepsen_ssh_key") + "\n"): ssh_auth_sock = os.environ["SSH_AUTH_SOCK"] auth_sock_dir = os.path.dirname(ssh_auth_sock) cmd = get_run_command( ssh_auth_sock, auth_sock_dir, pr_info, nodes_path, REPO_COPY, build_url, result_path, docker_image, ) logging.info("Going to run jepsen: %s", cmd) run_log_path = os.path.join(TEMP_PATH, "runlog.log") with TeePopen(cmd, run_log_path) as process: retcode = process.wait() if retcode == 0: logging.info("Run successfully") else: logging.info("Run failed") status = "success" description = "No invalid analysis found ヽ(‘ー`)ノ" jepsen_log_path = os.path.join(result_path, "jepsen_run_all_tests.log") additional_data = [] try: test_result = _parse_jepsen_output(jepsen_log_path) if any(r[1] == "FAIL" for r in test_result): status = "failure" description = "Found invalid analysis (ノಥ益ಥ)ノ ┻━┻" compress_fast( os.path.join(result_path, "store"), os.path.join(result_path, "jepsen_store.tar.gz"), ) additional_data.append(os.path.join(result_path, "jepsen_store.tar.gz")) except Exception as ex: print("Exception", ex) status = "failure" description = "No Jepsen output log" test_result = [("No Jepsen output log", "FAIL")] s3_helper = S3Helper("https://s3.amazonaws.com") report_url = upload_results( s3_helper, pr_info.number, pr_info.sha, test_result, [run_log_path] + additional_data, CHECK_NAME, ) print(f"::notice ::Report url: {report_url}") post_commit_status(gh, pr_info.sha, CHECK_NAME, description, status, report_url) ch_helper = ClickHouseHelper() prepared_events = prepare_tests_results_for_clickhouse( pr_info, test_result, status, stopwatch.duration_seconds, stopwatch.start_time_str, report_url, CHECK_NAME, ) ch_helper.insert_events_into(db="default", table="checks", events=prepared_events) clear_autoscaling_group()