The three-member team was too busy to burn down the storage hardware, and Microsoft admitted that its Azure campus was "understaffed."

Thanks to CTOnews.com netizen Coje_He for the clue delivery! CTOnews.com, September 5 (Xinhua) Microsoft's Azure Australian data center campus experienced a massive failure last week, when a decline in power supply capacity in the data center led to power outages in two data halls, which eventually led to the burning of some storage hardware and complete disruption of customer service for large enterprises, including the Bank of Queensland of Australia and Jetstar Airlines.

The service interruption report released by ▲ Jetstar Airlines at that time admitted that there were not enough staff at the scene and that its emergency procedures did not take into account the scale of the problem. At the time of the accident, only three staff members were assigned to the data center, all of whom dutifully "carried out the documented emergency operation procedures (EOP), but did not succeed in the end."

The ▲ source Microsoft's report on the incident (PIR) CTOnews.com found that the cooling capacity of the two affected data halls "consists of seven coolers, of which five coolers are in operation and two coolers are in standby state (Numb2)". The power supply sag (voltage sag) caused the failure of five running coolers.

Microsoft said, "the understaffing of the campus led to the accident, and the number of staff is now temporarily increased from three to seven before 'better understanding of the underlying problems and appropriate mitigation measures'." Microsoft will also adjust the relevant contingency plans and daily work processes to reduce the possibility of such incidents happening again.

▲ Picture Source Microsoft's report on the incident (PIR) it is reported that the accident affected seven large enterprise customers using Azure, including five at "standard" level and two at "advanced" level. Microsoft said that the hardware loaded with enterprise customer storage content was "damaged due to the high temperature in the data hall", except for heavy losses for enterprise customers. More than 250000 SQL databases hosted by the data center also failed, and according to Microsoft PIR, the relevant data recovery work has been largely completed.

▲ source Microsoft report on events (PIR)