China IDC Circle reported on August 21 that St The town of Ghislaina recently suffered from strong lightning weather, which also made Google's local data center unfortunately "lie down".
It is reported that the lightning weather hit a local power facility four times in total, leading to the power supply interruption of the main power system, and Google's data center was just located near the power facility. As a result, about 0.000001% of the data in Google's data center server has been permanently deleted.
Google said that the lightning weather at that time caused power interruption, resulting in disk damage, disconnection of some cloud storage systems, and data loss. Although the data center quickly switched to the standby power supply, this switching still resulted in the deletion of 0.000001% of the data and could not be recovered.
According to the news, Google's local data center is mainly used to drive Google Compute Engine services. This engine is mainly used for business, so it includes Google Mail Consumer services such as Youtube and Google Drive have not been affected.
In this regard, Charles David, head of a French start-up called "Azendoo", said that his company's services were interrupted for up to 12 hours because Google's data center was hit by lightning.
"Google successfully helped us recover some of the data, but a large part of the data still needs us to recover manually. Fortunately, Azendoo backed up its data in another Google data center, so we were able to retrieve all the lost data later. " David said.
In the subsequent official accident report, Google said that it needed to take full responsibility for the incident, and encouraged affected enterprise users to consider backing up data to other Google storage services.
"The accident of Google's cloud computing engine shows that enterprises who store all data in a single data center will inevitably face huge risks when encountering data center level accidents," Google said in its official statement.
As for Google's statement, Aaron Trubic, partner of Prop Villains, also agreed.
"Although this incident does not seem serious, it clearly shows the huge risks faced by enterprises to store all their data in a single data center," Tubick said.
It is reported that Google has launched the upgrading and improvement of data center infrastructure, and said that ensuring data security is the top priority of the company's current work. At present, most of Google's local servers have used storage hardware with better disaster resistance. In addition, Google also said that the computing architecture of the company's cloud services is all over the world, and users can set to automatically switch to the standby server when a disaster occurs. Among them, Google Cloud Computing Engine has three servers in the world, located in Belgium, Changhua, Taiwan, and Iowa in the United States.
"We have carried out a detailed analysis of this incident and are working hard to improve and enhance the reliability of Google's cloud computing engine. We also sincerely apologize to the users who were affected by this incident," Google said in a statement.
However, Joe Beda, a member of Google's cloud computing engine department, believes that although the company has suffered data loss in this incident, its rapid response ability is still worthy of affirmation.
"Google's data center suffered a major accident, but only 0.000001% of the data was lost. The Google Cloud Computing Engine team should be proud of this," said Beida.
However, although the proportion of data loss of 0.000001% seems very small, it may not be the case for Google, which has massive data. According to industry estimates, Google has about 10-15 exabytes of data (each EB equals 1 million TB), so 0.000001% of data is still equivalent to more than 100 GB of data content.
