Hacker News Reader

Comment by detaro | Hacker News Reader

This all being in one location with little redundancy ... somehow not surprised, but still, really?

Details are sketchy at this moment, but Aer Lingus runs lots of AS400 and Mainframe, MVS/ESA, IMS, DB2 and all that stuff. From the linked description on one of the articles

"...construction work damaged a fiber optic cable, wiping out the carrier’s check-in and boarding systems. Issues compounded after it was found that a network card used in the airline’s backup systems was not functioning, leaving Aer Lingus without passenger information for up to ten hours..."

So this looks like they had a backup location, but the new Kyndryl ( IBM ) team had not tested failover in a while... They are probably just trying to recover standard contract fees from Kyndryl specified in their SLA's

Edit: They sent this message at the time to their customers

https://img2.thejournal.ie/inline/5868785/original/?width=63...

> A network card used in the carrier's backup systems was not working

It should have been redundant, the problem is that the redundancy didn't work and I guess nobody noticed until it was actually needed.

It's the homeopathic backup problem — they make sure to have a backup because they know that's best practice, but they don't have any checks in place to make sure the backup works.

To be fair to the IT people, the failover system couldn't have been any more redundant.

> [IBM]

Anything the customer doesn't think to ask for, the customer doesn't get.

Reply on news.ycombinator.com