PayPal hardware failure fingered for worldwide outage
A brace of Friday fubars
Posted in Management, 1st November 2010 09:57 GMT
Watch Now : Virtual Machine Movement with Hyper-V
Hardware failure hit PayPal on Friday, leaving the payment service unavailable for around two and half hours.
A network hardware failure in one of its data centres and the failure of back-up systems to kick in as quickly as planned meant the service was unavailable for around two hours on Friday morning (US Pacific time).
A secondary problem meant the service was only restored for around two hours before going titsup again for an hour or so early in the afternoon.
PayPal CTO Scott Guilfoyle apologised for the outage and said it was investigating how to prevent any future re-occurrence of similar problems.
The problems at consumer-focused PayPal cap a miserable week for payment processors worldwide. Snags with the upgrade of systems at business-focused payment provider Sage Pay left merchants in the UK and Ireland struggling to complete purchases last Monday. ®
Watch Now : Virtual Machine Movement with Hyper-V
COMMENTS
"it was investigating how to prevent any future re-occurrence of similar problems."
Most likely it was a sneak attack by Cryptome and all those other account holders it shafted by freezing their accounts.
Users can cure the problem promptly, user a reputable service. Even AmEx is in on the business and they are experts at shifting money, especially circumventing currency controls for valued customers.
DR plan
Yeah, I have that to look forward to next week. Should be fun as..a sigmoidoscopy
- Test it and have it work in a real situation
And, of course, that carefully crafted test can never actually allow for what really happens at the worst possible moment.
Well,
backup contingency plans can only be properly tested if they're stress-tested. i.e. in the event of the contingency occurring for which they're designed.
If they're found wanting, then think again.
See multitudinous outages experienced by virtually every bank, every ATM network, every POS system etc. in recent times.
We live in a world reliant on complex real-time systems.
We also live in a world where shit happens.
It is no sin to make a mistake: it IS a sin to repeat the same mistake...
Paris? - just 'cause I mentioned sinning...
Its one thing to put in a DR plan in place.
Its another to actually test it.
Yeah... no joke. To test a D/R plan, you have to set up and schedule the down time and then have the risk if it fails and something then goes wrong.
IT infrastructure monitoring strategies
The new Office Garage series:
Top 10 SIEM implementer’s checklist
