Fix Continuous Integration #292
Conversation
We sleep 15 seconds to wait for GitLab to finish starting up, but because the sleep is outside of any test case / context, it happen to be executed _before_ manifests to install GitLab are actually applied. Move the sleep at the end of the first step of the test suite.
smortex
force-pushed
the
fix-ci
branch
2 times, most recently
from
c16571f to
b6fdac8
Compare
- Add scheme (http) - Use 127.0.0.1 to address the node itself, not 0.0.0.0 This should fix: ``` ubuntu1604-64-1 06:39:31$ /bin/sh -c curl\ -s\ -S\ 0.0.0.0:80/users/sign_in curl: (3) URL using bad/illegal format or missing URL ```
This may have helped diagnose the real source of the problems regarding CI.
|
@smortex ia ora na. Thank you for this PR. It looks that only puppet6-nightly fail. I don't understand why. I'll make some checks locally as soon as possible. |
|
@smortex this PR is interesting with In the last push, job in Travis failed because of true fail, probably linked to network issue or something like this. I asked to people with admin power in travis to rerun manually this uniq failing job. It should pass. |
|
@Dan33l sure. I can also force commit or squash commits so that Travis re-schedule a build, but it takes about 2 hours to run a build… I am not sure the sleep really helps here: as far as I can see, the build failures where the same when the sleep was running too early. Maybe it can be removed? |
|
Where it was previously in the code, the |
|
For sure it does not harm, we can leave it as it is now. 👍 |
Pull Request (PR) description
Travis CI is currently broken. Fix some issues found while inspecting the situation and attempting to unbreak CI.
This Pull Request (PR) fixes the following issues
No actual issues fixed, but it should help making #290 and #291 pass.