Hard Drive Temperatures

Below are graphs representing the temperatures of my RAID5 hard drives. I am conducting an experiment to determine if the Chenbro SR107 SATA hotswap backplane will keep my drives sufficiently cool. The WD400YR hard drives are in the hotswap backplane. The WD7500AAKS drives are in a much more open Lian Li PCA16B. Temperature data is collected by running smartctl every five minutes (i.e. using S.M.A.R.T. data).

Update: Saturday, October 27, 2007. I removed the hotswap backplane; I determined that it was simply too airflow restrictive to keep the drives sufficienctly cool without an obscenely loud fan. Now I am back to using the stock drive bays. The stock drive bays actually allow for roughly one-third an inch between drives. Also, the stock bays allow the use of an additional 92mm fan in front.


WD400YR (Four 400 GB Western Digital RE2)

wd400yr_temps
data_wd400yr_temps
  • Fri Sep 7 13:27:18 CDT 2007

    Initial testing began. Unfortunately, I didn't make a note of what fans I was using. My best guess is three Yate-Loon D12SL-12 fans (one in the back of the case, and two behind the drive bays). Clearly the drives are too hot. While they are within spec, they are simply too close to the max operating temperature for me to feel comfortable. Also, the 7500AAKS drives seem to run warmer by about five degrees Celsius; that would put them near the 60 degrees Celsius mark!

    During this time, I was also trying to stress the drives by running bonnie++ and updatedb.

  • Sat Sep 8 15:21:43 CDT 2007

    Rear case fan replaced with Antec tri-speed on high. It's kind of hard to see, as the graph is starting to compress, but this actually made a few degrees' difference.

  • Mon Sep 10 20:28:23 CDT 2007
    New fans:
    • Delta 120x38 Tri-blade in back
    • Scythe "F" on top
    • Delta 120x25 on bottom

    Note the dramatic swings in temperature around this time. This was due to leaving the windows open in the office over night. The ambient temperature in the room got pretty chilly. This is a major flaw in my testing methodology, by the way---failure to keep accurate ambient temperature measurements.

  • Mon Sep 17 20:02:30 CDT 2007

    Added aluminum tape in strategic locations, hoping to force airflow through hot swap bay

data_wd400yr_temps
  • Thu Oct 4 21:45:52 CDT 2007

    Installed high speed delta tri-blade in back

    Taped over more holes in back

data_wd400yr_temps
  • Thu Oct 18 19:07:51 CDT 2007

    Swapped /dev/sda and /dev/sdc

data_wd400yr_temps
  • Sat Oct 27 14:11:30 CDT 2007

    Removed hotswap, using stock drive bay

    92mm fan running in front of drives

  • Mon Nov 5 20:38:51 CST 2007

    92mm fan disabled

data_wd400yr_temps
  • Tue Nov 27 20:26:35 CST 2007

    Put (four) drives in SuperMicro CSE-M35T-1B SATA 5-in-3 backplane

    Swapped stock fan for delta AFB0912L Rev A before loading drives

  • Thu Nov 29 20:53:49 CST 2007

    Taped over some holes in Supermicro backplane

    Swapped fans to delta EFB0912L tri-blade (whiny)

  • Thu Nov 29 21:33:59 CST 2007

    Put AFB0912L Rev A back in---EFB0912L too whiny

 

WD7500AAKS - PCA16B

wd7500aaks_temps

Combined Data

combined_temps