Re: June 2007 computer failure
In June 2007 while the Orbiter Atlantis was docked at the ISS during mission STS-133, all 3 of the Russian segment GNC computers crashed. This left the ISS without the ability to do any propulsive maneuvers, had not the Orbiter been docked.
STS-117 was installing new solar arrays and the Russians were quick to blame power supply problems related to this for the failures. However, investigation showed that the actual problem was poor design of the computer system (which resulted in a zero fault tolerant situation) and their cooling systems (which led to corrosion issues).
In the weeks that followed the crisis and apparent recovery, station
commander Fyodor Yurchikhin and his fellow cosmonaut Oleg Kotov
disassembled the boxes and cabling and inspected every angle of the
hardware, occasionally assisted by their American crewmate, Clayton
Anderson. Multiple scopes and probes had failed to find the flaw, but
their eyes and fingers eventually did.
The connection pins from the power-monitoring device they'd bypassed
earlier, they found, were wet--and corroded. The final report
described the ”change in appearance” of fasteners on one box's
connectors and noted ”the presence of deposits and residue on the
housings, and residue and spots on the contact surfaces.”
Continuity checks found that specific wires, called command lines, in
the cable coming out of the device had failed. And one of those lines
had short-circuited. Also, in a shocking design flaw, there was a
”power off” command leading to all three of the supposedly redundant
processing units. The line was designed to protect the main computers,
which are downstream of the power monitor, from power glitches too
great for normal power filters to protect against. It does so by
turning the computers off when it senses trouble. But in a failure
unanticipated by its designers, this one command path itself was able
to kill all three processing units due to a single corrosion-induced
short.
During the next shuttle mission
During the August shuttle visit, the Russians were able to turn
stabilization control over to the American spaceship and tear down
their old computer network. The boxes and cables were replaced with
fresh units, built and supplied by the European Space Agency and sent
up inside a recently launched robot supply ship.
This quote
Multiple scopes and probes had failed to find the flaw, but
their eyes and fingers eventually did.
lends credence to the claim that recovery required astronauts on board. Also, the design error which resulted in a loss of redundancy / zero fault tolerance would have kept the ground from simply switching to a different computer.
Note, however, that the ISS was not designed to operate uncrewed for long periods of time, so it is unrealistic to expect it to be able to respond to failures without onboard intervention. See this question and its answers: Why has the ISS not been left unmanned?
Source