Airlines gradually came back online on Saturday after global airlines, banks and financial institutions were thrown into disarray by one of the largest IT crashes in recent years, caused by an antivirus update.

Passengers were crowded at airports Friday, waiting for news. Dozens of flights were canceled and airlines struggled to keep flights on schedule after an update to a program running on Microsoft Windows crashed systems around the world.

Several U.S. airlines and airports in Asia said they were now resuming operations. Check-in services have been restored in Hong Kong, South Korea and Thailand. Flights in India, Indonesia and Singapore’s Changi Airport were also largely back to normal as of Saturday afternoon.

“The check-in systems are back to normal (at Thailand’s five major airports). There are no long queues at the airports like we experienced yesterday,” Keerati Kitmanawat, president of Airports of Thailand, told reporters at Don Mueang Airport in Bangkok.

According to Microsoft, the issue started at 7:00 p.m. GMT on Thursday and affected Windows users running its CrowdStrike Falcon cybersecurity software.

CrowdStrike said a fix had been rolled out for the issue, and the company’s CEO George Kurtz told CNBC he wanted to “personally apologize to every organization, group, and individual that was impacted.”

It was also said that it may take a few days for things to return to normal.

U.S. President Joe Biden’s team has spoken to CrowdStrike and those affected by the outage and “stands ready to provide assistance as needed,” the White House said in a statement.

“We believe flights have resumed across the country, although there is still some congestion,” a senior U.S. government official said.

Reports from the Netherlands and the UK suggest that health services may have been affected by the disruption, so the full impact may not yet be known.

Media companies also faced problems. For example, the British Sky News reported that the outage caused news broadcasts to come to a standstill on Friday morning. The Australian broadcaster ABC also reported major problems.

Flights in Australia were largely back to normal on Saturday, but Sydney Airport continued to report delays.

Australian authorities warned of a surge in scams and phishing attempts following the outage, including people offering to help reboot their computers and asking for personal information or credit card details.

Banks in Kenya and Ukraine reported problems with their digital services, while some mobile operators saw service disruptions and a number of companies experienced customer service outages.

“The scale of this power outage is unprecedented and will undoubtedly go down in history,” said Junade Ali of the British Institute of Engineering and Technology, adding that the last incident of the same scale occurred in 2017.

– Flight chaos –

At some airports, all flights were halted, while at others, staff had to check in passengers manually, leading to long queues and frustrated travellers.

The US Federal Aviation Administration (FAA) initially ordered all flights grounded, “regardless of destination.” However, airlines later said they were restarting flights and clearing the backlog.

India’s largest airline Indigo said in a statement on X that the operations had been “resolved”.

“Although the outage has been resolved and our systems are back online, we are working hard to resume normal operations. We expect this to last into the weekend,” the airline said on Saturday.

A passenger told AFP the situation at Delhi airport had returned to normal by midnight on Saturday, with only minor delays to international flights.

Low-cost carrier AirAsia said it was still trying to get back online and was “working around the clock to restore its departure control systems (DCS)” following the global outage. It advised passengers to arrive at airports early and be ready for “manual check-in” at the airline’s counters.

According to Chinese state media, Beijing’s airports were not affected.

In Europe, major airports including Berlin said flights that were postponed earlier on Friday were resuming.

– ‘Common cause’ –

Companies had to repair their systems and assess the damage, while authorities tried to quell panic by ruling out foul play.

CrowdStrike’s Kurtz said in a statement that his teams were “fully mobilized” to assist affected customers and that “a fix has been implemented.”

But Oli Buckley, a professor at Britain’s Loughborough University, was one of many experts who doubted the ease of rolling out a good solution.

“Experienced users can implement the workaround, but it is not practical to expect millions of people to do so,” he said.

Other experts say the incident should prompt a rethink of the extent to which societies rely on a handful of technology companies for such a range of services.

“We need to be aware that such software can be a common cause of multi-system failures,” said John McDermid, a professor at York University in Britain.

He said infrastructure should be designed to withstand such common problems.