Flippant references to ''redundancy" make me cringe. They're usually written by people with a midrange view of the world, not a big data view. There's redundancy, then there's discovering bottlenecks you didn't know existed once your redundancy kicks in.
Just don't make it sound so much like "omfg didn't they have mirrored disks".
reliability vs. big data
Trading Systems like the ASX aren't doing big data, they are doing high volume, high value OLTP - that's a significant difference. Running such workloads would justify mission-critical systems, and in the past such systems were typically used by exchanges. But then came the big move towards plain vanilla systems, and nobody should be surprised when reliability suffers. Just adding HA clustering and mirrored disks isn't always sufficient ...
Shouldn't have used those spare machines from the Census guys
Epic snafu alright.
A number of instruments lost all record of last traded price data, before open this morning (the day after) there was a zero value for these instrument values i.e. books had a zero value on these, requiring a full query. This may point to no direct failover to a redundant system taking place and certainly no topend seamless 'failover'.
ASX CEO earning a topend fixed remuneration component of $2m+(STI+LTI variable components up to $5m total) & group execs including a CFO, Chief Risk Officer, Deputy CEO, Chief Information Officer, Chief Compliance Officer, Group Executive Operations exec all earning in excess of $650k I understand. I would think execs might try for a topend 'failover' as well, I know of some good people on here that would get it in shape.
ASIC today on users of the market :
"They have a legitimate expectation that Australia's securities markets perform to the highest standards."
Downvotes on El Reg just mean you've offended the ponytail penguin lovers. People who haven't worked with big BCP won't understand ASX's situation and will fall back to their level of comprehension "how come they didn't have another server ooooh"