I noticed something when i did a factory reset, and went from my "long term stable" setup with 4x2TB drives to a 4x500GB mixed models WD drive setup. And i have no idea what caused it, but it leads me to believe that the fan speed is software controlled. For an hour or so while it was testing the drives, the RAIDar software reported the NV+ fan speed as 3191 RPM, but ever since the space came on-line it's been sitting pretty at 2343 RPM. Now, these smaller drives are apparently hotter then my larger drives (most likely due to the fact that these are old recovered system drives, so 7200 RPM flavor where as the 2TB drives were 5400 RPM flavor storage drives.) and, i'd like to get the fan to push slightly more air if possible. I know this isn't really a priority, but, since the 4.1.15 firmware is "due" any time now, could we perhaps get a option to set the target temp for the sensor that is used to control the fan, or something?
It doesn't. And never did. With changing room temperature it goes up and down slightly, maintaining the same temperature (38-39c, the drives are at 41-45c) but calibrating the fan does nothing for me. I belive it's following a set profile, and i'd like to get it replaced, or to be able to tweak it. B!
I am not sure if this forum link below is applicable to your concern... However, I just want to share it to you: viewtopic.php?f=24&t=11799 - you may follow the steps indicated at your own risk...
http://www.readynas.com/forum/viewtopic.php?f=24&t=11799 - you may follow the steps indicated at your own risk...
As best as i can tell, the only thing they talk about "doing" is reversing the fan, providing worse cooling for the drives, but better for the PSU. This was since the "old" setting for the fan was 1600RPM or so on the bottom end, as opposed to our 2000-2100 currently. The problem i have isn't the PSU overheating, it's the drives i want cooler, preferably all below 40c. The addon they recommend could have been the solution, but if memory serves me right, all it did was set the bottom rpm to 2100, rather then letting you play with the value. And since i can't locate it, i cant verify this. The link to the thread where it's supposedly posted is easy enough to fix, just replace infrant, with readynas.
And... Yes, i know, 41-45 isn't considered out of line for most harddrives. Specs for the drives even states 0-60c as working temperature, so it isn't the end of the world. But i would sleep better knowing it's below 40, it's an old rule of thumb i've used for ages, and it eases my mind. I even kept "deathstar" IBM's alive for years while everyone else had them failing left and right. B!
Edit Thank you guys for pitching in tho. The effort is appreciated. /Edit
Google's drive whitepaper said that 30-40C was ideal, and ~25-45C had a minimal increase in failure rates. Every 5C that you go in either direction, the worse it gets.
I suspect it has something to do with the difference to ambient temperature, perhaps causing expansion/contraction. Since you're not in a cooled datacentre, your room temperature is likely higher, and 5C higher is probably fine.
Interesting. It's too bad that they didn't perform longer term testing, list the age of the drives, or list particular models. Other than indicating that the servers had been in use for a while, we have little to go on. It seems like 3-4 years might be the range, which does coincide with that huge spike in Google's whitepaper. (which clearly shows 30-35C as being most ideal) In that case, if they had performed the testing 3 years prior, it would've had a completely different conclusion. I don't think this one was long enough. I suspect a truly conclusive study will have to monitor drives for a period of a decade to get real insights into how and when they are afflicted with failures.
I suspect a truly conclusive study will have to monitor drives for a period of a decade to get real insights into how and when they are afflicted with failures.
Part of the puzzle is that drive technology changes during the study
I suspect a truly conclusive study will have to monitor drives for a period of a decade to get real insights into how and when they are afflicted with failures.
Part of the puzzle is that drive technology changes during the study
Too true. Several times, in fact. If you go back 10 years, perpendicular recording was the new thing displacing old methods. Now we've got shingled magnetic recording and Helium drives. Failure criteria could be different.
Testing older drives wouldn't be as useful these days. If someone started testing the new type (perp drives) the very day that they came out... didn't that include the 7200.10/11 generation? Oy! Firmware issues.
Still would be handy to see a breakdown of models tested.
FWIW, there is a plugin dedicated to the Sparc based redaynas (NV+) which allows to modify the min speedfan (was deigned to increase airflow towards PSU in earlier designs : EnableFanMinRPMOverride.
I found 2 ways to make the fan noise nearly absent:
#1 Upgrade to WD Red drives - not only have a had 0 failures with any of my Red drives (I have over 20 in various systems), but they are also very quiet and hover between 31C and 34C throughout the year (Bay Area, California).
#2 Upgrade the fan - my fan started acting flakey several years ago so I replaced it with the Noctua NF-B9. It was a fairly simple upgrade and one that made a noticeable difference in the noise factor.
Now I can have the system running 24/7 in the office with no complaints from the wife. I have a Windows Home Server for client backups and a newer NAS for data storage due to the much higher performance of modern systems, but the NV functions as the backup unit for those two servers. Considering I purchased the ReadyNAS NV almost 10 years ago, that is a powerful statement. Sure I upgraded the RAM to 1GB when the original 256MB stick failed, and I have upgraded the drives over the years (WD RE2 to RE3 to Red). Also the power supply failed and I replaced it with the much-improved fanless model, and the system fan needed replacement as I mentioned earlier. Other than those maintenance items, the NV has been running strong and provided with continual firmware upgrades as well; though I gather 4.1.15 will likely be the final one.
I didn't purchase it. The ReadyNAS came with a power supply that had a tiny fan that was very loud. Within a couple years of ownership the power supply failed and Infrant sent me a replacement power supply that had better components and no fan (totally silent). I'm assuming the power supplies may have been defective (or at least my batch was), because I recall this affected others on the forum. I haven't had an issue since, other than the RAM being replaced under warranty, and then the fan failing years later. I've also never had a drive fail, but I've always used enterprise drives that featured vibration protection; also I replaced the drives every 3-4 years as my capacity needs increased.
