Recently I have encountered many hard drive related errors on my home server. These errors were intermittent. Sometimes everything would be okay for a week, sometimes just a couple of days, and more recently within a few minutes. A sample of these errors would look like:
Aug 29 03:55:20 avs kernel: [47250.550847] ata9: softreset failed (1st FIS failed)I performed a SMART test on the hard drive and the tests reported that the drive was healthy, and yet the drive continues to fail.
I thought it was my SATA controller card, so I swapped the SATA port with another drive, and the same drive had the same issue. Therefore I knew it was not the SATA port, and the SATA cable was fine. I finally gave in and purchased a replacement drive. When the replacement drive arrived, I got the same error!
This is becoming extremely frustrating. The only thing that was left was the SATA power cable. I sourced the SATA power cable to another power source adapter, and finally the error went away. If I only knew!
Now I have an extra 10 TB drive. I didn’t want to return it, so I decided to buy another 10 TB drive and replace all the 4 TB drives in RAID 1 configuration that I had on the server. When I was replacing the drives, I noticed that one of the case fans was not working. It just so happens that this defective fan was sharing the same problematic power line.
After buying a $10 fan from CanadaComputers, and replacing the damaged fan, the server is now running without any issues along with the unintended storage expansion. I noticed from my UPS, the server is now consuming 10 W less.
In conclusion, the failed case fan probably caused some power anomaly. The fan did not totally fail because its LED was still working, but the fan motor seized and is no longer turning. This probably caused the extra 10W draw as the system was probably trying to overpower the unintended friction.
I’m documenting this because I find it fascinating that a simple failed fan could cause such instability to the system. Something to keep in mind for sure. Last point I want to make is that this fan is probably 15 years old, and absolutely more than a decade, so can’t really complain.