The consistency of a club-guessing failure at the successor of a regular cardinal

I answer a question of Shelah by showing that if $\k$ is a regular cardinal such that $2^{{<}\k}=\k$, then there is a ${<}\k$--closed partial order preserving cofinalities and forcing that for every club--sequence $\la C_\d\mid \d\in \k^+\cap\cf(\k)\ra$ with $\ot(C_\d)=\k$ for all $\d$ there is a club $D\sub\k^+$ such that $\{\a<\k\mid \{C_\d(\a+1), C_\d(\a+2)\}\sub D\}$ is bounded for every $\d$. This forcing is built as an iteration with ${<}\k$--supports and with symmetric systems of submodels as side conditions.