-
Notifications
You must be signed in to change notification settings - Fork 2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
subprocess.check_call + Rebar in Cowboy and Elli #495
Comments
One additional thing, I changed setup_util.py (which is called in line 6) so that the file write is both unbuffered and is explicitly closed. No effect. |
For reference: Round 7 preview data |
I haven't been able to reproduce the issue on EC2:
Results here: http://pastebin.com/L50nyiVE |
Thanks for taking a look at this @tcrossland. I'll run this by @ashawnbandy to see what we can make of it. |
Just to confirm, I haven't been able to reproduce the issue with elli either. For what it's worth, if the problem reoccurs, I'd suggest enabling verbose output on rebar
Good luck and thanks for putting these benchmarks together! |
@tcrossland Thanks so much for taking a look! We'll see if the verbose mode helps and if we can figure this out quickly. If not, we'll defer this until after Round 7 since it may be something very specific to our hardware environment. |
@tcrossland and @bhauer I certainly appreciate you taking a look at this as well - it's a puzzle I would love to see solved because I couldn't even get my finger in the crack to open it up. To add a couple of comments to what I wrote above:
|
Another idea. I just noticed the rebar binary in the framework tests is a little old (7 months). It might be worth modifying the install scripts to build rebar from the latest source: https://github.com/rebar/rebar. |
Fixed in master - See travis. Neither test passes, but it seems to be due to syntax errors and not anything systematic |
When run-test hits either the Cowboy or Elli test, it will either hang or time-out (depending on the version of benchmarker.py being used). run-test calls setup_erlang.py#start(args) and successfully executes all lines up to line 9 (see commit #4219cb3441). Line 9 is apparently called but never returns. This may be an intermittent issue but has persisted for at least the last seven days across several installations of the test harness and also appeared while @michaelhixson was testing a few weeks ago.
Additional facts:
The text was updated successfully, but these errors were encountered: