UK heatwave sparks cooling system meltdown in Google’s and Oracle’s London datacentre regions

The record-breaking heat that blighted the UK this week caused Google’s and Oracle’s local datacentre regions to experience technical difficulties on Tuesday 19 July, it has emerged.

On the same day the UK recorded the hottest day on record, two of the cloud market’s biggest players experienced cooling-related issues within their datacentres.

Google’s London-based region Europe-West2 experienced a cooling-related failure that began at 6pm on 19 July and was resolved, according to the company’s service status page, at 4am today.

“There was a cooling-related failure in one of our buildings that hosts a portion of capacity for zone Europe-West2-a for region Europe-West2 that is now resolved,” the service status page said.

The incident affected a “small set” of Google Cloud customers, the service status page continued, who would have experienced “abnormal [virtual machine] terminations”. Users were advised during the outage to work around this issue by relaunching their workloads in alternative European regions.

Meanwhile, the cooling infrastructure in Oracle’s UK South (London) datacentre was also adversely affected by the high temperatures, which forced the firm to power down a subset of its infrastructure to protect against uncontrolled hardware failures, its status page confirmed.

“This step has been taken with the intention of limiting the potential for any long-term impact to our customers,” it said.

During the outage, Oracle users were unable to access or use any Oracle Cloud Infrastructure resources hosted in the affected region, it confirmed.

Oracle explicitly cited the “unseasonal temperatures” as the cause of its issues, which are known to have begun just after 1pm on Tuesday and were resolved by 4am today.

Computer Weekly contacted Oracle and Google for comment on this story but both parties said they had no more to add at this time. 

The period of extreme heat seen in the UK this week has been linked to the onset of climate change, which is something datacentre resiliency think-tank the Uptime Institute has been warning datacentre operators to be mindful of for several years.

On this point, the Institute released research in 2018 that raised concerns about how few operators were factoring in the growing risk posed by climate change-induced freak weather events to their disaster recovery plans.

Adam Bradshaw, commercial director at Stevenage-based colocation provider ServerChoice, said the news that Google’s and Oracle’s datacentre infrastructure had suffered in the heat goes to show how even the industry’s biggest players are not immune to the effects of extreme weather.   

“It is not just people that have felt the effects of the scorching weather,” he said. “Datacentres that rely on cooling systems to keep their servers running have been experiencing difficulties too. While they are built to withstand all sorts of circumstances, many of these systems have been pushed to maximum capacity due to the temperatures we’ve experienced in this record-breaking week.

“The Google and Oracle outages as a result of overheating show that even the largest datacentre providers can fall victim to extreme weather.”

And with the sector already feeling the pinch from rising energy prices, the additional cost of keeping facilities cool during heatwaves may be too much for some to bear.

Bradshaw added: “As energy bills are a datacentre’s largest overhead, the rising price of energy has been hugely challenging to many businesses, even forcing some into administration. The added costs caused by the heatwave, and the likelihood that these temperatures will reoccur in years to come, highlights the need to alleviate the pressure on datacentres now and protect the interests of the country’s IT estate in the future.”