Astronauts restart computer
BY WILLIAM HARWOOD
STORY WRITTEN FOR CBS NEWS "SPACE PLACE" & USED WITH PERMISSION
Posted: July 15, 2011;
Adding discussion of shuttle flight rules for general purpose computer failures
Atlantis commander Christopher Ferguson and pilot Douglas Hurley carried out troubleshooting procedures early Friday and successfully restarted a shuttle flight computer that failed Thursday. But it is not yet clear what caused the initial failure and given the critical nature of the machines, flight controllers planned a thorough data review to make sure GPC-4 is, in fact, operating normally.
While Ferguson and Hurley focused on computer troubleshooting, the rest of the shuttle-station crew pressed ahead with a full slate of work to transfer supplies, equipment and other hardware to and from the International Space Station. The astronauts planned to take a break later in the morning, holding a traditional in-flight news conference at 9:24 a.m. EDT (GMT-4), and taking a call from President Barack Obama at 12:29 p.m.
The Atlantis astronauts are halfway through the 135th and final shuttle mission. Ferguson said the crew has been focused on the tasks at hand, but the reality of the looming end of shuttle operations is beginning to sink in.
"Of course, we've got a busy timeline and there's not a whole lot of time to think about it, but we realize as we go from module to module here, I mean literally we've reached the point where we're saying OK, it's the last train out of town, is there anything else that needs to go back to the Kennedy Space Center, back to Houston before the shuttle program wraps up here?
"And when you say things like that, I think the reality of it really begins to sink in, this is really the last train out of town. Of course, we're going to stay focused and very busy up until we undock, but you know, it's beginning to sink in. I don't think the full magnitude of everything is really going to hit us until after the wheels stop (on the runway)."
With thousands of shuttle workers facing layoffs, the mood at NASA field centers is somber at best, not so much because the shuttle program is ending but because funding was not provided to jump start development of follow-on spacecraft and missions.
As it now stands, it will be several years at best before U.S. rockets and new manned spacecraft are available to launch American astronauts. Until then, NASA will be forced to buy seats on Russian Soyuz spacecraft to get U.S. astronauts to and from the International Space Station.
But Ferguson is optimistic, looking forward to a new era of commercial spaceflight and eventual deep space missions using new NASA spacecraft and heavy-lift rockets that are little more than drawing-board concepts right now.
"Everybody realizes, the shuttle program, everything they've worked on for the last 30 years is coming to a close but by and large, we have something right behind it," he said. "And I think once we can finally get over the fact that the shuttle is gone and its day has come, I believe we'll begin to pick up the pieces and everyone will see that we really do have some vibrant programs out there that we're working on. So I know it's tough, it's kind of one of those wake-type things, you have to come to terms with the end before you can really put on a new beginning."
The glitch with general purpose computer No. 4 is the only significant technical problem encountered by Atlantis' crew since blastoff last Friday. While a second computer -- GPC-3 -- shut down during the shuttle's approach to the space station Sunday, flight controllers said a temperamental switch that briefly interrupted power to the machine was the culprit, not the computer itself. The crew restarted GPC-3 the next day and the computer is believed to be operating normally.
The shuttle is equipped with five IBM AP-101 general purpose computers. Four of them run identical software while the fifth runs programming from a different vendor to protect against bugs that might take down the four machines in the "redundant set."
During non-critical orbital operations, the backup flight system computer typically is kept in standby mode, as are two of the redundant set computers. Two GPCs are active, including one that is responsible for systems management. That is the one that failed Thursday, triggering an alarm that woke the crew.
Ferguson loaded the management software into GPC-2 and then went back to bed, leaving the shuttle under the control of GPCs 1 and 2 with GPC-3 in standby mode, ready for use if needed. Early Friday, he re-loaded GPC-4's software from a mass memory unit.
"And the data processing systems officer here in mission control reports that GPC-4 is now up and running in the common set," NASA commentator Rob Navias reported after an initial program load, or IPL, was carried out. "It appears, at least, that the recovery procedure has proven successful. If this computer remains up and running, it will be considered a transient failure and likely will be placed in a standby mode.
"The flight control team will watch the operation of this GPC for some period of time, but based on the way the procedures dictate, this GPC, although at the moment operable once again, would be considered a transient failure."
A transient glitch is defined as a single failure in a computer that subsequently responded normally to an initial program load. If it is determined GPC-4 was the victim of transient glitch and is otherwise healthy, it will be added to the redundant set for entry but it will not be used with any critical guidance, navigation and control systems.
Shuttle flight director Kwatsi Alibaruho said shuttle flight rules call for shortening a mission if two general purpose computers fail and cannot be recovered.
"If I lost two computers, then we would (shorten) the mission into what we call an MDF, or minimum duration flight," he said. "If I lost three computers, then according to my rules I would be looking at coming home essentially at my next primary landing site, essentially within the day. So as long as I have a computer that I have demonstrated I can load and bring up into our redundant set ... I don't have any acute mission duration impacts.
"Now, having said that, if the crew were to experience another problem with a different computer, honestly, I think we would all be thinking very, very hard about what was going on. We'd do what we always do, which is we would get data first, we would do data dumps of the computer that was functional and monitoring the other computers in the set to see what it saw, we'd do data umps of the failed computer to see what it saw about itself and we'd get the engineers to analyze that an take a good hard look at it."
Even with the failure of GPC-3, "which was due to a condition we have an explanation for, and the problem that we've had on GPC-4, which we don't yet have an explanation for, what I can tell you is that we have demonstrated the ability to bring both of those computers up to full computational readiness and add them to our redundant set," Alibaruho said. "So that's a meaningful capability we have."
Asked if he viewed the computer shutdown as a serious problem, a minor issue or something in between, he said "I consider every issue a major issue until we fully understand it."
"Right now, we've analyzed only part of the dumped data and we haven't seen any obvious software problems," he said. "Honestly, that's the thing I tend to worry about the most, some error in the flight software that we haven't caught. ... That's the scenario that tends to keep me up at night. Once we complete our analysis of the dumped data and determine we don't see any indications of that type of scenario, then I will be a bit more comfortable."
MISSION STATUS CENTER