Well looks like we both have own experienses, im currently trying to reproduce what happened to my scenario, so far no being able to have checksum errors, funny huh =P (could it be that something else have happened when fan was removed further away, i am open to ideas what have happened) (i do know that same have happened somebody else too, but why it has happened is not clear)
Fan vibration, not believe that, write/reading head thrown off by magnetfield no its so tiny and moves fast dont think so, heat not that too fan clearly kept hdd cool, memory no ecc mem had both cases, only what i can came up is that strong magnet because its a really really thin layer where data is stored on the platter.
Really i am open for ideas what it might have happened, this bothers me. (Yes i do know that case is aluminium alloy mixture and i know that aluminium dont do nothing on magneticfield, just tryed myself on opened case magnet did grab a nail just fine thought, as that dude in video demonstrated, also know that voicecoil magnets are 2 so field is really small) But if you dont have suggestions what have happened then next thing would this be good idea to give on mythbusters =) Still believe that magnet does superparamagnetic effect on platter. Not sure was those hdd models system area on edge of platter, if that happened it would mess hdd pretty good.
Please do tell those aluminium useful things of yours what it does than keeps body strong and spreads heat and reduces resonations.
Finally found what looked for i was lookin in wrong place, smartctl tells whats going on hdd before zfs kicks in data integrity.
(So looks like that smart tests are really really important to do if you want to know hows your hdd doing.)