There is a similar question herehere. It is asked in the particular case of the derived category of abelian groups and a counterexample is given by Tyler Lawson.
There is also an answer of mine which allows one to produce such triangles more generally.