Csdrhrt

Computer hardware question:

I installed an extra SATA hard drive in my Ubuntu 16.04 server a few days ago, to serve as a place to hold large datasets. It’s not configured as a boot drive, and it “works ok”, but I'm finding that machine will now occasionally freeze or turn off "randomly".

I have no idea why, or how to debug or fix this. Anybody have any suggestions?

• 
  EDIT: Now with the HD removed and everything set back to way things were previously, the system will STILL freeze. Ran memory checks, everything passed. Any ideas on what else to try?

Additional system info:

• The power supply on the server is "overpowered" - 850W capacity and we were drawing around 650 before adding the extra HDD.


• Two days ago, I found the machine had turned itself off. I turned it back on, and didn't notice any problems until...


• Last night it froze. The power to machine stayed on but it became unresponsive (no video output, pressing num lock didn’t cause the keyboard light to come on), my ssh connection was severed and the machine was un-ping-able.


• Last night when it froze, I was 15 minutes into running a calculation of the kind I’ve been doing on that machine for many months, and the part of calculation that reads the data from the drive had already been completed for ~15 minutes, so the extra "data drive" wasn't even being accessed.


• I just now rebooted from last night's freeze, and looked in /var/syslog, but there's nothing there to suggest anything unusual, e.g. nothing about any kind of crash, kernel panic, or “bringing the system down”, etc. The last syslog entry from last night is that it was about to run an hourly cron job (and /etc/cron.hourly is empty).


• Temporary BIOS change: One other possible hint: When I installed the drive, I thought maybe I needed to add it in the BIOS config menu somehow. Turns out I didn't, but while there I noticed the BIOS had three different system settings -- "Energy Saver," "Normal," and "Performance". I had it on "Normal," but tried changing it to "Performance." After that, I found that the machine wouldn't start up. Then I turned it off & on again, and it started up, but said "Overclocking Failed!" So I went back and changed the BIOS setting back to "Normal" and left everything else unchanged. Now it boots, but as I say, sometimes it freezes or dies. (Which motherboard/BIOS/chipset do I have? I'm happy to reboot and check, and edit this post later.)


• Ran GSmartControl tests: Basic Self-test, Short Self-test, and Conveyance Self-test have all completed without any errors.


• The relevant line in /etc/fstab reads as /dev/sda /mnt/data auto nofail,nosuid. Is there some obvious error in that?

?