Slashdot Mirror
BSOD Issues On Deepwater Horizon
ctdownunder passes along this excerpt from a NY Times article about a rig worker's testimony concerning the April 20 accident at the Deepwater Horizon well:
"The emergency alarm on the Deepwater Horizon was not fully activated on the day the oil rig caught fire and exploded, triggering the massive spill in the Gulf of Mexico, a rig worker on Friday told a government panel investigating the accident. ... On Friday, Mr. Williams added several new details about the equipment on the vessel, testifying that another Transocean official turned a critical system for removing dangerous gas from the drilling shack to 'bypass mode.' When he questioned that decision, Mr. Williams said, he was reprimanded. ... Problems existed from the beginning of drilling the well, Mr. Williams said. For months, the computer system had been locking up, producing what the crew deemed the 'blue screen of death.' 'It would just turn blue,' he said. 'You’d have no data coming through.' Replacement hardware had been ordered but not yet installed by the time of the disaster, he said."
The article doesn't mention whether it was specifically a Windows BSOD, or just an error screen that happened to be blue.
There is no evidence that BSODs contributed to this disaster. What is know to have contributed is the cheap cement job, plugged pressure sensors on the blowout preventer, possible damage to the blowout preventer during drilling (rubber fragments observed), and using seawater instead of drilling mud. None of these were automated.
I found this episode of 60 minutes quite interesting:
http://www.cbsnews.com/video/watch/?id=6490509n&tag=api
Apparently, BP was putting on a lot of pressure to do things quickly, since they were running behind schedule and it was costing them money.
Specifically, on the day of the accident, there was an argument between representatives of Transocean and BP on how to close the well (in preparation for later exploitation by another ship). Transocean was in favor the slower, safer procedure. BP wanted things to be done more quickly. They did it the BP way, which was the point when the accident happened. So, according to this report, there were BP emplyes on the Deepwater Horizont, and they influenced the procedures by pressuring their subcontractors.
According to the report, several other things had to happen as well in order for things to go wrong so badly, but I would not so easily let BP of the hook.
I was watching the testimony and he stated that it was a Windows NT system and was constantly giving a BSOD. They had replaced and reimaged the HDD over and over but it still kept happening. There were new servers, workstations, etc standing by and waiting to be installed, but another problem creeped in. They were waiting for another ship to figure out a way to run the old software on the new machines. Once that other ship could get it working and document it, they would then do the replacement on their end. I'm guessing it was a Windows NT 4 workstation.
If it turns out these crashes are Windows blue screens, the media will jump all over Microsoft
Well, before all the Microsoft Haters pile on, according to this the Control System in place was something called Cameron Multiplex Control System, which I've also seen referred to as Cameron MUX and CAMITROL. I am not pretending to be an expert in these things, just thought I'd share what little Googling turned up.
In short, it looks pretty unlikely that there's going to be a red hot poker headed toward Redmond over this.
Protip for all you people saying "They could have reconfigured the routers, etc." (on Childs refusal to hand over passwords) - not so much.
Why? Because Childs had either disabled serial consoles, disabled password recovery, or configured devices to -never- save configuration, only to run in RAM.
Well, shit, you say, restore the config from backups. Guess what, SF owned no backups of the configuration files, or network maps. The only configuration files Childs kept were on his personal laptop, encrypted with a key known only to him, and configured such that his laptop was the only device capable of updating configs. Network maps? Same. Sitting on his personal laptop. Nowhere else.
The guy viewed SF's network as his personal playground, and believed no-one else worthy to take the reins of it - guess what, he had no authority to decide that, and when he got nicely obstinate about it, he crossed a fairly clear line in the sand.
Stop the martyred geek defending valiantly our security creed. It bares little resemblance to reality.
Out of curiosity, where are you getting your information? You mean the last time there was a spill in the Gulf?
The oil spill off the Australian coast in August last year would seem to counter your claim entirely. Their procedures were lax and probably are still lax because those procedures are expensive. I'm not sure why people are so quick to dismiss the Gulf spill as a series of calamitous events when there is a ton of evidence to the contrary. Maybe people just want cheap gas too much and are willing to ignore all evidence so they don't have to face any consequences.
The sad reality is that oil is spilled everyday, Shell spills every year as much oil as the Exxon Valdez disaster in Alaska. Of course they spill it in Nigeria so no one cares.
Halliburton did the cementing job. http://online.wsj.com/article/SB10001424052748703572504575214593564769072.html
http://www.huffingtonpost.com/2010/05/10/halliburton-deepwater-hor_n_570733.html