Closed dean1012 closed 5 years ago
Restarted here: https://travis-ci.org/dokku/dokku-daemon/jobs/600531622
So odd...
Travis passed on my final commit and it appears it also passed on the PR.
It seems to consistently fail now though here:
Preparing to unpack .../dokku_0.19.2_amd64.deb ...
[....] Reloading nginx configuration (via systemctl): nginx.service7 ok 8Error: keyfile '/root/.ssh/id_rsa.pub' not found.
you can enter a new keyfile path when you restart the installation.
dpkg: error processing archive /var/cache/apt/archives/dokku_0.19.2_amd64.deb (--unpack):
new dokku package pre-installation script subprocess returned error exit status 1
nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
nginx: configuration file /etc/nginx/nginx.conf test is successful
Errors were encountered while processing:
/var/cache/apt/archives/dokku_0.19.2_amd64.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)
Makefile:66: recipe for target 'setup-travis' failed
make: *** [setup-travis] Error 100
The command "sudo -E make setup-travis" failed and exited with 2 during .
Tomorrow i'll pull up an Ubuntu Bionic image and try the tests locally to verify everything. In the meantime, if you have any thoughts (or anyone else) i'm all ears.
So it seems that the last successful TravisCI run (with no code changes after that) was 8 days ago. That would coincide with the release of dokku version 0.19.0.
I am thinking that something happened with either dokku version 0.19.1 or 0.19.2 that is causing the installation error of dokku here.
I will bring up a fresh Ubuntu Bionic tomorrow and try installation of both 0.19.1 and 0.19.2 from a fresh system and try to determine if there is in fact an issue with dokku (and if there is, I will create an issue on that repo) or if it's something else entirely.
On a side note, I also identified a bug in the TravisCI configuration. Specifically, the dokku version environment variable is defined twice so in effect it always will be master.
I will create a pull request that removes the second DOKKU_VERSION variable and keeps the one that sets it to master.
Yeah its not clear to me why the keyfile stuff broke since I haven't touched that in a while, so worth checking that out I guess...
Tests pass now. Thanks!
It looks like tests on the master branch after merging #25 have failed due to what looks to be a fluke issue.
Can you please trigger a Travis run on the master branch manually to verify that tests actually do succeed @josegonzalez
If tests still fail, i'll need to investigate.