I have the monthly parity check plugin installed.  I usually run it while everyone is asleep, so I don't really listen to it.  Today, I was awake at 3 am, and I started listening to a loud humming sound coming from my server.  At first, I thought it was a fan going bad, but I check them and the sound was from one of my hard drives.

 

I couldn't tell, which one was making that sound though.  Is there a way to find out if a hard drive is going bad?  The parity check isn't done yet, and I saw these error in the log list:

 

Nov  1 02:49:31 Tower kernel: Pid: 2138, comm: unraidd Not tainted 3.4.11-unRAID #1 (Errors)

Nov  1 02:49:31 Tower kernel: Call Trace: (Errors)

Nov  1 02:49:31 Tower kernel:  [<c10551fc>] print_cpu_stall+0x59/0xd1 (Errors)

Nov  1 02:49:31 Tower kernel:  [<c10552af>] __rcu_pending+0x3b/0x125 (Errors)

Nov  1 02:49:31 Tower kernel:  [<c105540f>] rcu_check_callbacks+0x76/0xa1 (Errors)

Nov  1 02:49:31 Tower kernel:  [<c102b12b>] update_process_times+0x2d/0x58 (Errors)

Nov  1 02:49:31 Tower kernel:  [<c1048e6a>] tick_periodic+0x63/0x65 (Errors)

Nov  1 02:49:31 Tower kernel:  [<c1048e85>] tick_handle_periodic+0x19/0x6c (Errors)

Nov  1 02:49:31 Tower kernel:  [<c10173fa>] smp_apic_timer_interrupt+0x67/0x7a (Errors)

Nov  1 02:49:31 Tower kernel:  [<c131ff92>] apic_timer_interrupt+0x2a/0x30 (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84bb87e>] ? xor_sse_5+0xc8/0x3d8 [xor] (Errors)

Nov  1 02:49:31 Tower kernel:  [<c10161ab>] ? native_smp_send_reschedule+0x3f/0x41 (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84bc8de>] xor_blocks+0x66/0x71 [xor] (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84bc8de>] ? xor_blocks+0x66/0x71 [xor] (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84f292b>] check_parity+0x8d/0xc3 [md_mod] (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84f34b5>] handle_stripe+0xa7d/0xcfb [md_mod] (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84f37aa>] unraidd+0x77/0xb8 [md_mod] (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84f0925>] md_thread+0xcc/0xe3 [md_mod] (Errors)

Nov  1 02:49:31 Tower kernel:  [<c1035eac>] ? wake_up_bit+0x5b/0x5b (Errors)

Nov  1 02:49:31 Tower kernel:  [<f84f0859>] ? import_device+0x147/0x147 [md_mod] (Errors)

Nov  1 02:49:31 Tower kernel:  [<c1035bb4>] kthread+0x67/0x6c (Errors)

Nov  1 02:49:31 Tower kernel:  [<c1035b4d>] ? kthread_freezable_should_stop+0x49/0x49 (Errors)

Nov  1 02:49:31 Tower kernel:  [<c13208f6>] kernel_thread_helper+0x6/0xd (Errors)

Nov  1 04:10:38 Tower kernel: NTFS driver 2.1.30 [Flags: R/W MODULE]. (System)

 

I will attach the complete log list once the parity check is done.  I've never done hard drive smart check, but I will do those as well.  I am hoping to catch if a drive is going bad before it happens so I can be ready for it.  Thanks.

 

UPDATE:  Adde system log file.  No parity errors.

syslog-2012-11-01.txt

After the parity check is complete, go to the main page and spin all of the drives down.  Then spin them up one at a time and see if you can determine which one it is.  If you can determine which one it is, my decision would be easy--- REPLACE THE DRIVE!!!! Any unusual noise from a hard drive is not a indication that things are normal regardless of what any smart test or other data check says.  In cases of potential drive failures, early replacement is far, far easier than waiting for it to actually happen.

 

 

After the parity check is complete, go to the main page and spin all of the drives down.  Then spin them up one at a time and see if you can determine which one it is.  If you can determine which one it is, my decision would be easy--- REPLACE THE DRIVE!!!! Any unusual noise from a hard drive is not a indication that things are normal regardless of what any smart test or other data check says.  In cases of potential drive failures, early replacement is far, far easier than waiting for it to actually happen.

 

Yeah, that's what I was thinking.  I'm probably going to start buying 3TB drives, and hoping one of the other drives can replace the one that is bad.  I am also hoping it is only one drive.

Or you may be able to figure it out while the parity check is still going on with the "screwdriver ear".

 

Put the handle of a screwdriver in your ear and the point of the screwdriver on each drive. The source of the noise should be apparent.

Or you may be able to figure it out while the parity check is still going on with the "screwdriver ear".

 

Put the handle of a screwdriver in your ear and the point of the screwdriver on each drive. The source of the noise should be apparent.

 

I thought about that, but didn't do it.

 

 

UPDATE:  I spun up and down each drive individually, and no of them made any noise.  I could hear them spinning up and down.

Maybe the parity check was making a fan run and it was the source of the noise? Fans are usually the source of the louder noises in my experience.

Maybe the parity check was making a fan run and it was the source of the noise? Fans are usually the source of the louder noises in my experience.

 

I think it might be one of the exhaust fans.  I turned the server on, left it off for a few minutes.  I turned it back on, and I could hear like a bad bearing noise coming from the back top area of the case.  I'll replace the fans and see if that noise goes away.

whoooweeee,

I'd rather replace a fan than a drive any day.

whoooweeee,

I'd rather replace a fan than a drive any day.

 

Tell me about it.  Well, I can't remember the size for the exhaust fans for my Helios 910 case, and I'm trying to find some good quiet and inexpensive long lasting fans too!  Anyone?

 

I has been 2 days, and I haven't hear the same noise at all even if I spin up and spin down the drives.  Even the fans are quiet now, but I can hear the fans make a noise after a while.

I have had fans make intermittent noise so that is a definite possibility.  (One of mine was on startup when the fan was cold.)  The easiest way to check is to stick something (not your finger) into the blades of the fan while the noise is being made. If it stops, you have identified the problem.