In a bid to put its long-running check-in woes behind it, Virgin Blue will be switching off most of its computer systems from 8pm tonight until 5am on Thursday — meaning no online bookings or check-in, no phone sales, and no kiosk check-ins at airports. If you’re flying with Virgin Blue during that time, get yourself to the airport even earlier than usual; if you’re planning to fly, either book this afternoon or wait until later in the week. [Virgin Blue]
Lamby
Tuesday, October 5, 2010 at 4:58 PMAnother failed Microsoft real-time mission critical systems to go along with the London Stock exchange.
SQL and .NET strikes again
(http://www.microsoft.com/casestudies/Case_Study_Detail.aspx?CaseStudyID=200076)
http://blogs.computerworld.com/london_stock_exchange_to_abandon_failed_windows_platform
DavidRa
Tuesday, October 5, 2010 at 6:15 PMActually, Lamby, everything I’ve seen points to a failure in processes – not a failure in the technology stack. Yes, LSE was a big stuffup – but let’s make sure we don’t cast stones at the wrong parties.
http://www.theregister.co.uk/2010/09/28/virgin_blue/ suggests it was initially a storage failure – Texas Memory Systems (the RAMSAN is an extremely high performance SAN storage array) or NetApp. http://www.theregister.co.uk/2010/09/29/netapp_virgin_blue/ suggests it wasn’t the NetApp controllers. The same article then suggests it took 21 hours to decide to cut over to the DR hardware platform – and that’s a failure in process, not technology. If they’d cut to DR immediately (ie within 10 minutes) most people wouldn’t even have noticed except for slight delays in booking/checking in/boarding for that 10 minute period and however long it took to catch up again.
If you have a real time environment like this, and it takes 21 hours to decide you’ve had a disaster, your process is broken and you need to fix it. Technology cannot solve process problems; it can only move the problem elsewhere.