Ongoing power outage (2020-12-10, Amsterdam, power backplane defect

  • Friday, 11th December, 2020
  • 05:32am

This message has been sent to update you regarding the ongoing power outage affecting one or more of your services.

The worse case scenario has materialized. The power backplane of the enclosure blowing up has fried all blade servers in the enclosure. We are concluding this based on the fact that our spare blades are powering up normally and giving a video signal while in the enclosure.

It is unknown yet what the extent of the damages are. The HP SL230s blade servers themselves have a power distribution board (also called the "personality board"). This is a swappable card that connects into the power backplane of the enclosure. It is possible that only this component has been damaged. However, it is also possible that the damages extended to the mainboard and up to the remaining components (processor, memory, RAID controller and drives). There is a possibility at this point that the electrical damages extended to the drives and compromised your data.

Such a catastrophic failure wasn't anticipated. We've never ever had a power supply or power backplane blowup and damage other components. We were unaware that it was even a possibility for the damages to extend simultaneously to all 8 blade servers of the enclosure. As such, we chose to store only 2 spare blades on-site for the 24 total blades that are hosted at our Amsterdam location. We are therefor unable to provide an immediate solution to all 8 blade servers affected by this incident.

The next best troubleshooting step is to assess whether the damages extend beyond the personality board.

We have instructed the on-site staff to swap the personality board of one of the affected blade servers with the one from one of the spare blades.

If this fails to restore access to the affected blade server, then we will have the on-site staff move drives from one of the affected blades to a spare blade.

After these last troubleshooting steps, we will know the extent of the damages and what next steps is most appropriate.

The on-site staff has now departed the datacenter. We expect that there will be no availability for remote hands until tomorrow morning (local time).

Please also be advised that, in the event that swapping all personality boards is found to be the best solution, that there will be a very long lead time to source these parts.

In order to best serve you going forward, we will need the following information :

What operating system was installed on your affected servers and was the HP B320i RAID controller enabled?

This information is critical, as moving drives between HP SL230s units where the HP B320i RAID controller is enabled requires both units to have the controller enabled. The opposite is also true.

Thank you,

« Back