Can't restart travis-ci job

[nunnatsa]

When trying to re-run a travis job, it never succeed and reports the following error:

$ $GOPATH/bin/goveralls -service=travis-ci -coverprofile=#####
Bad response status from coveralls: 422
{"message":"service_job_id (####) must be unique for Travis Jobs not supplying a Coveralls Repo Token","error":true}
The command "$GOPATH/bin/goveralls -service=travis-ci -coverprofile=######.coverprofile" exited with 1.

[gdm85]

Have you set COVERALLS_SERVICE_JOB_ID or TRAVIS_JOB_ID?

[nbaztec]

If the travis job id doesn't change on re-run (which I think it doesn't), this can definitely cause the error on coveralls side, since the job_id needs to be unique per push without repo-token or omitted. @gdm85 I think we need a mitigation strategy for this scenario in general.

[gdm85]

What does coveralls.io suggest to do in this case? Travis CI is officially supported by them.

[nbaztec]

Seems to be an issue on coveralls' side.

Can be closed.

[gdm85]

Let's close it until it can be reproduced again and with a better context.