The nova issue here is the conductor retries of 3 is too small for medium scale (64 node) clusters. While I realize its configurable, my expectation is OpenStack services work out of the box at 100 node scale under overload conditions.
The nova issue here is the conductor retries of 3 is too small for medium scale (64 node) clusters. While I realize its configurable, my expectation is OpenStack services work out of the box at 100 node scale under overload conditions.