Viraj's Weblog

We have a Coraid device with 10 120gig drives setup as storage for our mailserver. This Saturday at about 5am the Coraid powered off, and appeared to be a problem with the power supply. I came in later in the morning to check it out.

We took apart the unit and found that it has a simple rackmount ATX power supply. The motherboard connector plugs into the Coraid backplane and 3 HD power connectors also plugin to the same board.

We had an ATX power supply tester which plugs in to the motherboard connector and has an on/off switch. We connected it to the power supply, switched it on, and it powered on. Then we connected it back to the Coraid, and the unit powered on as well. We put it back together and it seems stable.

Obviously we'd like to prevent this from recurring. Coraid support told me that they've never heard of a problem like this. They mentioned it might have been a power glitch causing the switching power supply to freeze. They asked us to check our UPS load, but it is very low and the other systems on it exhibited no problems. They claimed that it does not sound like a defective power supply.

I'm not so sure. Honestly these boxes should be built with redundant power supplies, but unfortunately it seems rather cheaply made. We have to keep our eye on the unit, but we still don't know why it failed. We will probably buy another power supply just to be safe.