How hot can a harddisk become???
 

While reading the syslog I noticed a pretty high temperature for my harddisks:

How hot can a harddisk become???

Carefully feel the drive. Either the smart sensors is in error or the drive is on its way out. Most WD idle at 26 and peak at 50. Seagates tend to run at 33 and peak at 60.

Thanks for the warning, but I did not feel the heat.

I think that the smart sensor is too optimistic (or pessimistic).

IF it thinks that the temp is in C but it is really in F the temperature would be in about the right range. (50C=122F, 43C=109F)

111 C has to be wrong. That is significantly above the boiling point of water, and I would expect your drive to go into thermal shutdown by then. If it really is that hot, you'll know it as soon as your hand gets close to it. You'll likely get burned if you touch it.

Try installing smartctl and/or sensors on your system to see what you get from those.

My SCSI IBM Ultrastars (15K RPM drives) run between 33 and 40C. My SCSI Fujitsu (15K RPM drive) runs at around 27C - it barely gets above room temperature. My Seagate SCSI drive (7200 RPM) runs around 35. My old Quantum SCSI drive (7200 RPM and makes as much noise as the rest of the system combined) runs around 35C. Of course, I have fans blowing on all of them to keep them cooled down, but this should be indicative. The room this workstation sits in is kept at around 24-25C.

With Suse smartd is default installed. I agree with you that the disk would die at 111 C.
Tested it with my hand but the temperature is quite healthy, so I guess that the sensor is nuts.
With smartctl no problems were detected.

2008/08/15

I concur with the earlier posted advice (above), it looks like the temperature is being mis-reported as Celsius instead of Fahrenheit.

I work, and have worked with, a lot of well used HDDs of different types. As a rule of thumb HDDs can run fairly warm to the touch, but if they feel more than very warm then there is a problem. The problem stems either from the HDD assembly itself, or more often from a lack of adequate airflow over and around the HDD.

Most HDDs are fairly robust (manufacturing defects aside) and can operate for long periods while very warm, but ultimately an HDD that operates at higher temperatures for long periods will be prone to early failure. (Lookup the manufactures specifications for the normal operating temperature range)

In my experience, one of the most common items neglected in PC assembly is the provision of adequate airflow about a HDD. Closely followed by the obstruction of airflow within the case by numerous cables.

May I suggest that if you are worried, check that the HDDs in your system are well spaced, replace any flat data cables with round equivalents and group them together using cable ties, out of the path of the primary air flow in the case, and consider investing in an additional air intake fan on the bottom of a side panel in the case to increase the efficiency of the PSU's exhaust fan.

Third party Heatsinks for HDDs still exist, but are less common these days.

Hope that helps

Regards

Chris

PS. For a better indication of the external temperature of the HDD, you could use a thermocouple attached to a suitable multimeter. I recommend this method when evaluating the ambient temperature inside the case.

C

This must be a bug in the smartd reporting.
(Reporting in F instead of C)

My logs show pretty much exactly the same messages
on the last 2 machines I have put Fedora 8 on.

Install hddtemp to get the right numbers.

Here are the last 2 lines from my current messages file, and the
output from hddtemp: