Timer relay fault probably caused power cut to part of GMO Internet data centre in Japan
Jan 2016Published 13 Oct 2026Responsible editor: Chief Editor
Get the next one by email
Free alerts on data centre incidents in Japan, and an email when EIR-2353 is updated. No account needed; unsubscribe in one click.
Get a free sample brief Three complete verified records and the incidents-per-GW table, by email within a minute.
By signing up you agree to the privacy policy.
Companies named in this entry: GMO Internet
What happened
On 16 January 2016, at about 14:15, power was lost to roughly 12% of a GMO Internet data centre in Japan. Several brief power interruptions occurred first, and power returned automatically after each one. GMO attributed the outage to a partial failure of power equipment in the data centre. Its provisional finding was that a timer relay (time-delay relay) in a transformer and distribution panel had malfunctioned, most likely causing the loss of supply.
GMO first investigated the location it judged most likely, which delayed identification of the root cause. Services hosted across several data centres were affected, including the domain registration service Onamae.com, rental servers, ConoHa by GMO and GMO AppCloud. Some customers could not display their control panels. Service was fully restored at about 14:25 on 17 January 2016, roughly 24 hours later.
Contributing factors
Multiple brief interruptions made the fault pattern hard to read, so the investigation began with the location judged most likely, which delayed identification of the root cause. Customers' services were spread across several data centres, and not all customers of the same service were affected. The detailed cause was still under investigation at the time of GMO's statement.
Lessons learned
- Test and maintain time-delay relays and other protection components in transformer distribution panels, because one small component can remove supply to part of a facility.
- Where a fault produces a series of transient interruptions, review the full event log before focusing on one location, so that diagnosis is not delayed.
- Spread services across independent data centres and power supply paths so that one fault does not affect every customer of a service, and publish clear, timely status updates during an outage.
More lessons on electrical protection and isolation, spacing, layout and fire spread, from every incident on the register.
Source
Published 2026-10-13. Written from public reporting. Descriptive, not investigative. See the data accuracy disclaimer.
