Recommended hard drive temperature
in [Storage]
|
From: Franc Zabkar on 21 Apr 2008 01:40 On 17 Apr 2008 20:56:34 GMT, Arno Wagner <me(a)privacy.net> put finger to keyboard and composed: >Seek errors are due to modern drives >starting reading before the heads have settled. This usually works, >but when it does not work it becomes a seek error. Some >manufacuters list these in the SMART data, other do not. The >number seen does not mean much, which is well known to people >that work a lot with SMART data. > >Arno I demonstrated elsewhere in another NG that in Seagate's case the "seek error rate" figure is actually a count, not a rate, and it is a count of the total number of seeks, not seek errors. I did this by performing a zero fill operation on a 13GB drive and recording the SMART "seek error rate" parameter before and after. See ... http://groups.google.com/group/microsoft.public.windowsxp.hardware/msg/2ac63d875bfaf0d4 .... for my results. - Franc Zabkar -- Please remove one 'i' from my address when replying by email.
From: Arno Wagner on 21 Apr 2008 05:39 Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: > On 17 Apr 2008 20:56:34 GMT, Arno Wagner <me(a)privacy.net> put finger > to keyboard and composed: >>Seek errors are due to modern drives >>starting reading before the heads have settled. This usually works, >>but when it does not work it becomes a seek error. Some >>manufacuters list these in the SMART data, other do not. The >>number seen does not mean much, which is well known to people >>that work a lot with SMART data. >> >>Arno > I demonstrated elsewhere in another NG that in Seagate's case the > "seek error rate" figure is actually a count, not a rate, and it is a > count of the total number of seeks, not seek errors. I did this by > performing a zero fill operation on a 13GB drive and recording the > SMART "seek error rate" parameter before and after. > See ... > http://groups.google.com/group/microsoft.public.windowsxp.hardware/msg/2ac63d875bfaf0d4 > ... for my results. Ah, yes. Bottom line, the "Seek Error" Attribute is pretty meaningless, if you do not know the specific drive. Arno
From: Arno Wagner on 21 Apr 2008 05:48 Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: > On 20 Apr 2008 22:03:24 GMT, Arno Wagner <me(a)privacy.net> put finger > to keyboard and composed: >>Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: >>> On Sun, 20 Apr 2008 14:16:28 +0200, lars <lars(a)hesdorf.dk> put finger >>> to keyboard and composed: >> >>>>In short, time well spend reading. >>>>http://www.pdl.cmu.edu/PDL-FTP/Failure/failure-fast07_abs.html >> >>> This document appears to be a statistical analysis of HD failures. It >>> doesn't attempt to delve into the technical reasons for failure. The >>> only time it discusses temperature, or SMART, is in reference to the >>> Google article in my OP. >> >>> Google's experience suggests to me that temperatures below about 35C >>> result in greater failure rates, which is contrary to normal >>> expectations. However, Arno appears to be saying that the lower temps >>> may be a consequence of failure rather than a cause. >> >>Exactly. It is possible, but the paper does not give us enough >>data to determine whether it is the case. Also it runns contrary >>to all known reliability characteristics of semiconductors, >>other electronics components and mechnanics. >> >>Arno > What about fluid dynamics? Maybe there is an optimal temperature for > the platter lubricant and/or air bearing. Possibly. Many drives in the Google study should actually be pre-fluid bearing, if I remember correctly when they became mainstream. A part would be FDBs though and maybe there is some increased vibration effect or the like at lower temperaturers. Now what would be interesting is SMART status changes for the drives that dies at lower temperatures, compard to those that died at other temperatures. Also temperature vs. FDB percentace would be of interesst and temperature vs. disk age would be too. Also disk performance in the week before death vs. temperature would be nice. > I found this interesting Samsung patent whose inventors claim that > "flying height drops significantly in humid conditions" and that this > can be remedied "by increasing the temperature of the air flowing > between a slider's air bearing surface and the rotating disk surface > it accesses". > Method and Apparatus Reducing Flying Height Drop in a Hard Disk Drive > Under Humid Conditions: > http://tinyurl.com/4s5brl > http://www.freshpatents.com/Method-and-apparatus-reducing-flying-height-drop-in-a-hard-disk-drive-under-humid-conditions-dt20071227ptan20070297085.php Interesting, I will have a look at there references! Not relevant for data-center operation, however, since humididy is also strictly regulated in there. Arno
From: Franc Zabkar on 21 Apr 2008 16:51 On 21 Apr 2008 09:48:45 GMT, Arno Wagner <me(a)privacy.net> put finger to keyboard and composed: >Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: >> What about fluid dynamics? Maybe there is an optimal temperature for >> the platter lubricant and/or air bearing. > >Possibly. Many drives in the Google study should actually >be pre-fluid bearing, if I remember correctly when they became >mainstream. A part would be FDBs though and maybe there is some >increased vibration effect or the like at lower temperaturers. When I wrote "fluid dynamics", I was referring to the air flow under the R/W head, ie the air bearing, not the motor bearing. >Now what would be interesting is SMART status changes for the >drives that dies at lower temperatures, compard to those that >died at other temperatures. Also temperature vs. FDB percentace >would be of interesst and temperature vs. disk age would be too. >Also disk performance in the week before death vs. temperature >would be nice. > >> I found this interesting Samsung patent whose inventors claim that >> "flying height drops significantly in humid conditions" and that this >> can be remedied "by increasing the temperature of the air flowing >> between a slider's air bearing surface and the rotating disk surface >> it accesses". > >> Method and Apparatus Reducing Flying Height Drop in a Hard Disk Drive >> Under Humid Conditions: >> http://tinyurl.com/4s5brl >> http://www.freshpatents.com/Method-and-apparatus-reducing-flying-height-drop-in-a-hard-disk-drive-under-humid-conditions-dt20071227ptan20070297085.php > >Interesting, I will have a look at there references! Not relevant >for data-center operation, however, since humididy is also strictly >regulated in there. > >Arno Static electricity becomes an issue in low humidity environments. I recall one site where the Control Data hard drive would log a "status error" whenever the operator touched it. The solution was an antistatic mat. At other sites I've seen humidifiers used to solve this kind of problem. I would think that any datacenter with a humidifier would encounter the issues addressed in the Samsung patent. - Franc Zabkar -- Please remove one 'i' from my address when replying by email.
From: Arno Wagner on 21 Apr 2008 19:04
Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: > On 21 Apr 2008 09:48:45 GMT, Arno Wagner <me(a)privacy.net> put finger > to keyboard and composed: >>Previously Franc Zabkar <fzabkar(a)iinternode.on.net> wrote: >>> What about fluid dynamics? Maybe there is an optimal temperature for >>> the platter lubricant and/or air bearing. >> >>Possibly. Many drives in the Google study should actually >>be pre-fluid bearing, if I remember correctly when they became >>mainstream. A part would be FDBs though and maybe there is some >>increased vibration effect or the like at lower temperaturers. > When I wrote "fluid dynamics", I was referring to the air flow under > the R/W head, ie the air bearing, not the motor bearing. Ah. That would be a different type of dynamics, that, while having some fluid properties, is not fluid dynamics. >>Now what would be interesting is SMART status changes for the >>drives that dies at lower temperatures, compard to those that >>died at other temperatures. Also temperature vs. FDB percentace >>would be of interesst and temperature vs. disk age would be too. >>Also disk performance in the week before death vs. temperature >>would be nice. >> >>> I found this interesting Samsung patent whose inventors claim that >>> "flying height drops significantly in humid conditions" and that this >>> can be remedied "by increasing the temperature of the air flowing >>> between a slider's air bearing surface and the rotating disk surface >>> it accesses". >> >>> Method and Apparatus Reducing Flying Height Drop in a Hard Disk Drive >>> Under Humid Conditions: >>> http://tinyurl.com/4s5brl >>> http://www.freshpatents.com/Method-and-apparatus-reducing-flying-height-drop-in-a-hard-disk-drive-under-humid-conditions-dt20071227ptan20070297085.php >> >>Interesting, I will have a look at there references! Not relevant >>for data-center operation, however, since humididy is also strictly >>regulated in there. >> >>Arno > Static electricity becomes an issue in low humidity environments. I > recall one site where the Control Data hard drive would log a "status > error" whenever the operator touched it. The solution was an > antistatic mat. At other sites I've seen humidifiers used to solve > this kind of problem. I would think that any datacenter with a > humidifier would encounter the issues addressed in the Samsung patent. I am just saying that very likely all HDDs in the study were running with similar humidity, and therefore humidity will not be a factor examined. Arno |