search

strongloop / strong-pm

deployer for node applications
http://strong-pm.io
Other
1k stars 71 forks source link

Mark processes dead more robustly on startup #204

Closed sam-github closed 9 years ago

sam-github commented 9 years ago

connected to strongloop-internal/scrum-nodeops#571

sam-github commented 9 years ago

failed because log output didn't have "listening on"... but its hard to tell, in the DEBUG build, there are 100s of lines of output until the app output is seen. Could just be a race or a buffering issue.

I don't think its related, hopefully, the next test will pas.

@rmg, are you aware of any instability in that test?

# pmctl log using REST API
ok 280 cmd "status 1" pattern /Processes/
ok 281 exit status should be zero for: cmd "log-dump 1" pattern /.+ worker:1 pid \d+ listening on \d+/
not ok 282 cmd "log-dump 1" pattern /.+ worker:1 pid \d+ listening on \d+/

http://ci.strongloop.com/job/strong-pm.pr/601/label=amazon-64-0.12/consoleText

rmg commented 9 years ago

That test could fail if the test system was under heavy load and the worker hadn't actually started listening yet. Looking at the test app it looks like there are some semi-static strings printed immediately on startup that might make for better test strings.

rmg commented 9 years ago

LGTM.

rmg commented 9 years ago

@slnode test please

rmg commented 9 years ago

Rerunning CI now that #206 has been merged. Should pass now.

sam-github commented 9 years ago

@slnode test please