Apollo 6 (AS-502) was the second flight of the Saturn V (the first having been Apollo 4/AS-501, five months earlier), the last flight of a Block I CSM,1 and the last unmanned Apollo mission (if one doesn’t count Skylab 1, which was less of a “mission” and more of an “enormous cargo flight”).
It was also the Apollo mission that came closest to failing outright,2 suffering from large first-stage pogo oscillations3 which severely damaged the engines in the Saturn V’s second (S-II) and third (S-IVB) stages. This caused two of the S-II’s five engines to shut down prematurely4 (forcing the remaining three engines to burn to depletion to compensate; even then, an extra-long S-IVB burn was still necessary to make up the remaining difference), and caused the S-IVB’s single engine, following the prolonged insertion burn necessitated by the partial S-II failure, to fail to restart.
Any of these failures, had they occurred on a manned mission, would have required an immediate abort.
All later Saturn Vs were modified as a result, with changes to the S-II and S-IVB igniter plumbing to strengthen the fuel and oxygen lines against the pogo-induced ruptures that occurred on Apollo 6, and with parts of the first-stage (S-IC)’s LOX system charged with gaseous helium in an effort to “detune” the S-IC and damp out any oscillations that might occur. These modifications, however, were not flight-tested prior to the first manned Saturn V flight (Apollo 8, or AS-503 - as its numerical designation indicates, only the third-ever flight of a Saturn V), and the engineers had no way of knowing whether their anti-pogo measures would actually be effective under the stresses of an actual launch, or whether the pogo would return,5 or whether the modifications would themselves introduce new, unforeseen problems. Nowadays, if a set of failures this serious occurred during the final shakedown flight of a new human launcher, NASA would demand at least three or four more unmanned flights to test the resulting modifications and prove that the problems wouldn’t recur before even thinking about human-rating the vehicle.
Additionally, Apollo 8 would be the first time a Saturn V was used for a translunar flight, the first time a Saturn V carried a Block II CSM (and, consequently, the first flight of a fully-fuelled Block II CSM, as the Saturn IB used for Apollo 7 was incapable of lifting a fully-fuelled CSM into orbit), and, therefore, also the first translunar flight of a Block II CSM.6 The following manned Saturn V flight (Apollo 9, or AS-504) would be the first time a Saturn V ever carried an LM and the first time a CSM and LM were ever mounted and launched together, while the first translunar LM flight (Apollo 10, or AS-505) was, again, a manned mission.
Why was NASA okay with human-rating the Saturn V immediately after the serious failures experienced on Apollo 6, and despite the fact that this would result in astronauts being sent up with hardware, combinations of hardware, and mission profiles that had never been tested on an unmanned flight?
1: Albeit somewhat modified to test some Block II features.
2: ...if one doesn’t count Apollo 1, which didn’t fail, per se, but, rather, was cancelled following a fatal ground testing accident.
3: A type of harmful oscillation caused by coupling between, on the one hand, the bending modes of parts of the rocket’s structure, and, on the other, the variations in engine thrust that this bending causes by way of stretching and bending the engines’ fuel and oxidiser lines.
4: One of the engines shut down following a chamber failure in the engine’s igniter, caused by the pogo rupturing the igniter’s LH2 line (starving the igniter of fuel and allowing the chamber wall to be blowtorched by essentially pure oxygen); the second engine flamed out when a defect in the S-II’s engine control wiring allowed the shutdown command intended for the first engine to also close the fuel line for the second engine (which, until then, had been operating mostly-normally).
5: As it turned out, the problem had been essentially solved for the S-IC, but it turned up for the S-II, with a vengeance, on Apollo 13 (AS-508), where extreme pogo oscillations during second-stage flight nearly tore the Saturn V apart; only a momentary, and likely spurious, low-chamber-pressure signal from the center engine - resulting in the premature shutdown of said engine - saved the vehicle from destruction. (Drawing on this knowledge, all following Saturn V flights shut down the S-II center engine much earlier than the outboard engines, finally solving the pogo problem.)
6: For that matter, Apollo 7, the first manned Apollo mission that actually flew, was also the first-ever flight of a Block II CSM (of the six unmanned Apollo missions - four Saturn IB missions and two Saturn V missions - one Saturn IB mission [AS-203] was an S-IVB checkout flight with no CSM or LM, two Saturn IB missions [AS-201 and AS-202] flew with Block I, non-human-ratable CSMs, and one Saturn IB mission [Apollo 5] was an LM test flight with no CSM, while the two Saturn V missions [Apollo 4 and Apollo 6] also used Block I CSMs); while the Block I CSMs on the two unmanned Saturn V missions had both been modified to test some Block II features, no actual Block II CSM flew on an unmanned mission before its first manned flight with Apollo 7. At least Apollo 7 was a LEO flight, allowing a quick return should a problem have developed; the same was not true for Apollo 8 (or for any other of the translunar Saturn V flights), with even the quickest abort from a translunar trajectory being considerably longer than an abort from LEO.