libvirt CI failing due to Jenkins problems
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Linaro CI |
Triaged
|
Undecided
|
Riku Voipio | ||
Bug Description
** I was requested to submit a specific bug regarding the libvirt CI failures to make sure it is tracked correctly **
The libvirt CI job has been failing for 2+ months now due to problems with the CI environment.
Originally, the problem was specifically due to failures that were occurring during Jenkins steps to copy files between master and slave, i.e.:
http://
(from https:/
It was suggested that there were ways to restructure the build jobs in order to eliminate the need for these copies, but, there is not a way (without making changes to all the CI slaves) to eliminate the slave->master copy that is required for pushing .debs to repo due to the slaves not having SSH keys.
I was able to work around this failure for a while by grabbing the completed .debs from the build workspace.
However, the libvirt build has been failing for some tie now at a much earlier step in the build process, i.e.:
http://
(from https:/
Because of this earlier failure, it is not even possible to grab intermediate build output from the workspace.
| Riku Voipio (riku-voipio) wrote | #1 |
| Changed in linaro-ci: | |
| assignee: | nobody → Riku Voipio (riku-voipio) |
| Changed in linaro-ci: | |
| status: | New → Triaged |
| Chase Qi (chase-qi) wrote | #2 |
Builds are now succeeding again;
https:/
Notice that the packages are now being uploaded to trusty, so the rest of CI loop in lava side needs to be updated to install packages from trusty. I'll dive into that tomorrow.