Errors have been introduced into more than 19,000 National Australia Bank accounts as the bank worked to rectify a "corrupted file" that crashed its payment processing system on Wednesday night.
The four-hour outage delayed "all types of payments" - including Centrelink welfare, credit card, salary and interbank transactions - due into accounts on Thursday November 25.
Transactions made on Friday were also delayed as NAB worked through a backlog of potentially "millions" of accounts.
The bank today stated that it had fixed the glitch, and promised to "put things right" for any customers who had incurred costs from the delays.
However, "duplicate or missing transactions" were introduced in around 19,000 accounts "because the system was under some duress," a spokesman told iTnews this morning.
"Over the course of the weekend, we've managed to rectify the situation," the spokesman said, adding that NAB would rectify any remaining errors today, and overnight if necessary.
Besides delaying payments to and from NAB customer accounts, Wednesday's outage also affected customers of NAB's partner banks, HSBC and Citibank.
NAB processes direct entry and over-the-counter deposits for Citibank, as well as interbank transfers, payroll payments, direct debits, cheque payments and deposits for HSBC.
A Citibank spokesman told iTnews on Friday that it had been in contact with its corporate customers and was responding to retail customer queries about any processing delays.
Citibank anticipated that NAB would quickly rectify the technical issues that resulted from its "infrastructure failure".
HSBC apologised to customers for any inconvenience, urging those who needed immediate access to funds to visit an HSBC branch.
"We're working with NAB to resolve the issue as a matter of urgency and we hope to expedite customer payments as soon as possible," a HSBC spokesman said.
NAB said the outage was a one-off incident, but University of Sydney computer science professor Alan Fekete said sensible enterprise system design would have kept systems operational despite data corruption.
Data could have been stored in multiple places, to assist in recovery and keep systems running should one machine crash or one copy of data be corrupted, he said.
"It's possible to remain viable once a computer glitch kicks in by using a variety of hardware and by keeping records so you can untangle problems after the event," he stated.
"Proper data management stays operational and ensures an audit trail. In this case the system stopped operating when a file became corrupted, suggesting an oversight in the design of the wider IT systems."
Updated at 2.22pm to include Alan Fekete's comments.