Comment 2 for bug 1482195

Revision history for this message
Yanis Guenane (yanis-guenane) wrote :

The deployment itself was working fine on a local setup but kept failing on the CI.

This discarded an issue in the logic of the script/puppet.
After some investigation and some discussion/logs analysis in #clusterlabs, it came out that it was apparently a CPU issue (not enough CPU) causing corosync/totem to behave in a weird way.

Would this be possible to update the CI to use NODE_CPU=2 at least and either confirm or discard this hypothesis ?