This article claims it was common in the USSR if an enterprise needed steel, to buy say knives or other steel goods and melt them for steel.
In general, control by the ruble was designed to prevent deviations from the central production plan. But since the plan itself was often inconsistent, providing an enterprise with too little of one input and too much of another, managers in order to meet their output requirements were forced to develop sources of supply that could bypass Gosbank's clearing operations, i.e., sources that required neither bookkeeping money nor cash. Hence, the immense amount of interfirm bartering that took place in the Soviet Union. An enterprise with excess coal might be lucky enough to trade it for some desperately needed steel. More likely, it would trade its excess coal for some rubber that it didn't need, and would then go about finding an enterprise that had excess steel but needed rubber. Or, worse still: it would trade coal for rubber, then trade rubber for steel knives, and finally melt down the knives to obtain raw steel.
I wonder if this is correct.