Identify noisy disk(s) in 8-bay RAID
I have an OWC Thunderbay 8 filled with 8 Ironwolf Pro 10TB drives in a RAID 5 configuration. The drives are all new and SoftRAID reports no problems with the disks.
The problem is I'm getting 51.5 decibels measured at 2-inches from the front of my enclosure. It's utterly intolerable. Even placing the enclosure in a cabinet with sound insulation the noise is driving me to distraction.
I've written to Seagate to ask whether this is characteristic of the drives and the cumulative effect of 8 of them, or a mechanical problem with one or more. Based on this recording that I sent them, they suspect the latter: https://f.io/SaLPgP3G
Could anyone offer an opinion on this?
At 0 - 23 seconds, I'm clicking between video clips but not playing them. At 24 seconds, I click play on a video clip.
If this sounds wrong, can anyone suggest the best way to find the problem drives(s), ideally without breaking the software RAID/data that I have set up? Can I remove the drives and reinsert one-by-one to test, or will this mess things up?
I'm on Big Sur 11.1, SoftRAID 6.0.1 b48. 16 inch Intel MacBook Pro.
Thank you in advance,
Simplest thing to do is shut down. Remove all the disks slightly (1/2 inch is enough)
Power on the computer and drives. Insert one drive only. listen, then pull it out and let it spin down, then insert the next drive.
See if one of them is much louder than the others. If you only do one or 2 drives at a time, it won't affect your volume, as there are not enough disks to try to mount it.
Thank you for getting back to me. I subscribed for notifications to this thread but did not receive one, hence my slow response.
Your suggestion was helpful. No one disk is appreciably louder than the next when attached to the device and left to idle for 60 seconds. The test did not allow me to read/write which is where the real noise problems arise, but I'm guessing the result would be the same.
These disks really seem so much louder than the WD Reds that I've been used to in the past.
You can also try "verify disk" when you have one at a time. It may help.
Look up the specs, are the WD 5400 vs new disks 7200 RPM?
Otherwise, drives should be similar in noise.
That's it. The WD's are 5400 RPM. I think I need to adjust my expectations then, perhaps buying some working U2 drives for smaller projects where I don't need the 70TB RAID capacity and can turn it off.
Just a quick one on heat... Putting the drive into an insulated cupboard does reduce the noise, and I can go further with more soundproofing. Of course, it is getting a bit warm in there.
Am I right in thinking that SoftRAID does not measure disk temps? I believe I read that is because heat does not correlate strongly with higher failure rates? Are there any guidelines that suggest keeping ambient air temp below X degrees, for example?
Thanks for your help.
HDD's are always a noise generator. Depends on how small the cabinet is, I would be somewhat careful. If you can put a small fan to vent out the back, that would help. Drives should be below 40 degrees C, in general, but can handle 50 degrees. the google study did not find a correlation with temperature and heat. I am mixed on this, as temperature extremes obviously decrease the lifespan of any thing electronic or mechanical.
Anyway, that is why SoftRAID does not measure temperature. The main issue is "Why do you measure? To what specific standard ?" There is no set threshold of OK vs Bad, so it will lead to ambiguity and confusion to users. Just because a drive was exposed to 50 degrees does not mean it will die tomorrow. If at 50 all the time, however, it is likely to die sooner than one at 35 or 40 continuously.