2

It appears quite strange to me that temperature readings derived from SMART for one disk is different from its twin in a RAID 1 configuration by as much as 9°C:

# smartctl -d scsi -A /dev/sg1
=== START OF READ SMART DATA SECTION ===
Current Drive Temperature:     34 C
Drive Trip Temperature:        68 C

# smartctl -d scsi -A /dev/sg2
=== START OF READ SMART DATA SECTION ===
Current Drive Temperature:     43 C
Drive Trip Temperature:        68 C

The server is a 1U Dell PowerEdge R210 with the specs here. What could be the cause of such an anomaly?

6

This is far from abnormal. This is my server's HDD temperatures over the past year:

Annual munin graph of HDD temperatures

They are correlated, sure, but one runs about 7C hotter than the other. In my case I'm fairly sure it's down to airflow differences inside the case: one catches more of the flow from the case fan than the other. In the past, when I've swapped them round, the lines have simply swapped with them (those data have aged off the graph, sorry).

Google's famous paper about HDD failures concluded that

at moderate temperature ranges it is likely that there are other effects which affect failure rates much more strongly than temperatures do.

so as long as you're staying below 45C on a regular basis, it may not be worth worrying too much about the discrepancy.

  • Nice chart. By the way are the two horizontal lines your threshold values? If so, why are they not the same and how do you determine these values? – Question Overflow Jul 23 '14 at 13:09
  • Yes, they are alerting thresholds. They aren't the same because the two drives' characteristic temperatures aren't the same. I picked them because they tend to alert me when there are genuine issues - like the one you see at the far RHS of the graph, which had me spending most of last Saturday in a colo. You tune your monitoring systems until they give neither false positives nor false negatives, basically. – MadHatter Jul 23 '14 at 21:31
1

What is the configuration of the physical disks and/or fans near them?

It is entirely possible to see a wide range of change if the drives, are for example, on top of one another in a cramped case.

I'd look at the physical design for your answer.

  • I have updated my question with the specs. The drives are side by side, not stacked. – Question Overflow Jul 23 '14 at 5:07
  • You could look at the layout of the CPUs/fans relative to the disks to help explain the discrepancy. – dmourati Jul 23 '14 at 6:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.