Concur with Tony Wells, I build tube gear with high voltage power supplies, always use a bleeder resistor to bring down the charge in a capacitor. After discharge, I might short the capacitor just before removal to dissipate any memory/residual charge. The capacitors can hold a charge for a long time, so do not assume that since you haven't used a machine for awhile that the capacitors are discharged. When dealing with banks of capacitors, there is a tremendous amount of charge held in the capacitors unless there is some form of bleeder resistors across the terminals, and even then they can take some time to fully discharge. When I purchase old high voltage military oil can type capacitors, they always come with a wire across the terminals when packaged new.
From another source: Electrolytic capacitors may become permanently damaged by excessive peak currents with rapid discharge/charge, which will definitely occur during short-circuit events. The reason is that (a) the internal resistance will cause a momentary, but large power dissipation (heat!) and (b) the distribution of the current spike inside the capacitor will not be formed evenly across the large area of the aluminum foil and hot spots may occur. The electrolyte may vaporize along these small zones and damage to the insulating aluminum oxide layer may occur as well. If you're lucky, the capacitance will decrease just a bit or the top of the can may change its shape into something like a dome. If you're very unlucky, the cap may fail and heat up quite a bit (and eventually blow). This applies to different types of capacitors, not just electrolytic types.