Japan’s Kyoto University has lost some 77TB of research data due to an error in the backup system of its Hewlett-Packard supercomputer, it was announced today.
Today, due to a malfunction in the backup program of our storage system, there was an incident in which some files in /LARGE0 were lost. The process in question has been stopped, but we are investigating the extent of the impact as nearly 100TB of files may have been lost.
This incident occurred between December 14 and 16, 2021, and resulted in 34 million files from 14 research groups having their systems and backup files erased.
After studying the impact of the loss, the university has concluded that the work of four of the affected groups can no longer be restored.
All affected users have been notified individually via email, but details about the type of work lost have not been released.
The backup process is stopped at the moment.
To prevent a recurrence of data loss, the university plans to decommission the backup system, make improvements, and re-install it in January 2022.
We will also be doing incremental backups (for files that have changed since the last backup) as well as full backup mirrors.
Supercomputers are expensive
The details of the type of data lost are not known, but since supercomputer research can cost hundreds of US dollars per hour, the incident must have been distressing for the affected groups.
Kyoto University is one of the most important research institutions in Japan and the second largest university in terms of investment in scientific research funded by government grants.
The excellence and importance of its research is particularly evident in the field of chemistry, which is ranked fourth in the world, and also contributes to biology, pharmacology, immunology, materials science, and physics.
We have asked Kyoto University to explain the details of this incident and its impact on our research group, but have not yet received a response.
Japan leads the world
Japan is home to the most powerful supercomputer in the world today, called Fugaku, operated by the RIKEN Center for Computational Science in Kobe, Japan. https://www.hpcwire.com/off-the-wire/fugaku-retains-title-as-worlds-fastest-supercomputer/
“Fugaku” is an exascale system made by Fujitsu, with a computing performance of 442 PFLOPS.
IBM’s Summit, ranked second in the world, has 148 PFLOPS, which is much lower than Fugaku.
“Fugaku” was built at a cost of $1.2 billion and has been used for COVID-19 research, diagnosis, treatment, and simulation of the spread of the virus.
Comments