Today I tried to deploy a new service, changeprop. Scap3 was able to fetch and check out the code, but the promote stage failed without a clear reason why.
mobrovac@tin:/srv/deployment/changeprop/deploy$ deploy -v --force 18:06:13 Started deploy_changeprop/deploy 18:06:13 Update server info Entering 'src' 18:06:13 == CANARY == :* scb1001.eqiad.wmnet 18:06:13 Running remote deploy cmd ['/usr/bin/deploy-local', '-v', '--repo', 'changeprop/deploy', '--force', '-g', 'canary', 'fetch'] 18:06:13 Creating /srv/deployment/changeprop/deploy/.git/DEPLOY_HEAD deploy_changeprop/deploy_fetch: 100% (ok: 1; fail: 0; left: 0) 18:06:15 Running remote deploy cmd ['/usr/bin/deploy-local', '-v', '--repo', 'changeprop/deploy', '--force', '-g', 'canary', 'config_deploy'] 18:06:15 Creating /srv/deployment/changeprop/deploy/.git/DEPLOY_HEAD deploy_changeprop/deploy_config_deploy: 100% (ok: 1; fail: 0; left: 0) 18:06:15 Running remote deploy cmd ['/usr/bin/deploy-local', '-v', '--repo', 'changeprop/deploy', '--force', '-g', 'canary', 'promote'] 18:06:15 Creating /srv/deployment/changeprop/deploy/.git/DEPLOY_HEAD 18:06:15 ['/usr/bin/deploy-local', '-v', '--repo', 'changeprop/deploy', '--force', '-g', 'canary', 'promote'] on scb1001.eqiad.wmnet returned [70]: 18:06:15 INFO - Starting new HTTP connection (1): deployment.eqiad.wmnet deploy_changeprop/deploy_promote: 100% (ok: 0; fail: 1; left: 0) 18:06:15 1 targets had deploy errors Stage 'promote' failed on group 'canary'. Perform rollback? [y]: n 18:06:27 Finished deploy_changeprop/deploy (duration: 00m 13s)
What I gather here is that deploy-local is trying to connect to deployment.eqiad.wmnet. However, the service's scap.cfg says:
[global] git_repo: changeprop/deploy git_deploy_dir: /srv/deployment git_repo_user: deploy-service ssh_user: deploy-service server_groups: canary, default canary_dsh_targets: changeprop-canary dsh_targets: changeprop git_submodules: True service_name: changeprop service_port: 7272 lock_file: /tmp/scap.changeprop.lock [wmnet] git_server: tin.eqiad.wmnet