noisy (whining) IDE RAID disks in server

[xxlukexx]

Hi,

We have an old Compaq server running a Compaq smart array controller
431 attached to four IDE disks, in RAID1.

The disks have been getting noisier lately, as old disks tend to.
However, I came in today and the server was switched off. Upon starting
it up the disks are making a very loud whining, metal on metal scraping
noise - I mean much louder than any disk should be, and several times
louder than they were a week ago.

I've tried unplugging the power leads from the disks one by one to work
out which one is failing and making this noise, but I've found that 3
out of 4 are all doing it. I'm totally confused - how could 3 disks all
fail within a week of each other?

I've tried every combination of plugging the disks in or leaving them
unplugged and I'm certain that 3 are making this noise. What's going on
here? Is it something that the computer has done to the disks to make
them fail? I just can't believe that three would have broken, in the
same way, at the same time, when this server has been running without
missing a beat for three years.

One thing I've noticed - they don't make the noise when the server
starts booting. It's not until the array controller starts it's
intialisation sequence that the noise starts. Don't know if this is
because the disks don't get spun up until this point (I know normal IDE
disks spin up as soon as they get power, but these are all attached to
the controller) or whether the controller is doing something that's
made them fail?

Any help or ideas would be greatly appreciated.

Luke

 

[xxlukexx]

Just realised - this is a SCSI controller and disks, not IDE as I
thought.

 

[J. Clarke]

Hi,

We have an old Compaq server running a Compaq smart array controller
431 attached to four IDE disks, in RAID1.

The disks have been getting noisier lately, as old disks tend to.
However, I came in today and the server was switched off. Upon starting
it up the disks are making a very loud whining, metal on metal scraping
noise - I mean much louder than any disk should be, and several times
louder than they were a week ago.

I've tried unplugging the power leads from the disks one by one to work
out which one is failing and making this noise, but I've found that 3
out of 4 are all doing it. I'm totally confused - how could 3 disks all
fail within a week of each other?

I've tried every combination of plugging the disks in or leaving them
unplugged and I'm certain that 3 are making this noise. What's going on
here? Is it something that the computer has done to the disks to make
them fail? I just can't believe that three would have broken, in the
same way, at the same time, when this server has been running without
missing a beat for three years.

One thing I've noticed - they don't make the noise when the server
starts booting. It's not until the array controller starts it's
intialisation sequence that the noise starts. Don't know if this is
because the disks don't get spun up until this point (I know normal IDE
disks spin up as soon as they get power, but these are all attached to
the controller) or whether the controller is doing something that's
made them fail?

Any help or ideas would be greatly appreciated.

What you're describing sounds like a physical crash--having one happen on
three drives at once is rare though, unless there was an impact of some
kind--is there any possibility that the machine got knocked over or some
such?

In any case, I suspect that the three noisy drives are going to die rather
quickly.

 

[Arno Wagner]

(e-mail address removed) wrote:

What you're describing sounds like a physical crash--having one happen on
three drives at once is rare though, unless there was an impact of some
kind--is there any possibility that the machine got knocked over or some
such?

My thoughts exactly. Alternatively, may be overheating. Ball-bearings
are quite sensitive to heat. It may have driven them over the
edge. Maybe a cooling failure or heavy load some time in the recent
past.

In any case, I suspect that the three noisy drives are going to die
rather quickly.

I second that. Back up now and replace them. You should still
find out what the common cause is, since three doing this
within a week is _very_ unlikely and you may need to fix the
underlying issue as well.

Arno

 

[Folkert Rienstra]

What you're describing sounds like a physical crash--having one happen on
three drives at once is rare though, unless there was an impact of some
kind--is there any possibility that the machine got knocked over or some
such?

In any case, I suspect that the three noisy drives are going to die rather
quickly.

They already did if they make that kind of noise.
If they are still working nonetheless then there is no saying how long
they will be working because headcraches would kill them in an instant
and if that didn't happen then it aren't headcrashes causing the noise.

 

[xxlukexx]

Thanks, everyone, for all your comments. I've got some replacement
drives coming tomorrow, the only thing that concerns me is what caused
them to fail in the first place - i.e. is it going to happen to the new
drives. Only one way to find out, though.

The drives are still working, although I've had event log errors over
the weekend. Now that the machine has been running overnight and has
warmed up, the noise is not so bad as it was yesterday when I was
starting it up from cold.

It's all very weird - maybe these drives have been getting slowly
noisier and I haven't noticed, and booting from cold really brought
then noise out. Being a server they run all day every day, so have
probably "got used" to being warm to a certain

Hopefully replacement drives will sort it all out,

Cheers,

Luke

 

[Folkert Rienstra]

Thanks, everyone, for all your comments. I've got some replacement
drives coming tomorrow, the only thing that concerns me is what caused
them to fail in the first place - i.e. is it going to happen to the new
drives. Only one way to find out, though.

The drives are still working,

So it is very likely spindle bearing noise.

 

[Curious George]

My thoughts exactly. Alternatively, may be overheating. Ball-bearings
are quite sensitive to heat. It may have driven them over the
edge. Maybe a cooling failure or heavy load some time in the recent
past.


I second that. Back up now and replace them.

I'd sure hope he had regular backups running with some retention even
before any bad noises were heard.

You should still
find out what the common cause is, since three doing this
within a week is _very_ unlikely and you may need to fix the
underlying issue as well.

It's uncommon but not the craziest thing. These disks are likely the
same model, maybe if they came with the server even from the same lot,
installed at the same time, shipped the same, and have had identical
usage in close proximity in the same case. It's far more likely for
them to suffer very similar fates than in many other scenarios of
machines with 4 drives - although yes there is still a much stronger
possibility that they suffer from a common insult.

But unless someone knocked it over or the AC failed, etc. I'm not sure
there is any "underlying issue" he could/should undertake to fix (i.e.
engineered obsolescence). If the machine has been up for years
without problems until recently - and the only present issue is noise
coming from functioning but aging disks - that is really quite
respectable. Maybe it's me but once he replaces the disks, & checks
logs including environmental & server access - I don't think he really
has much more to do.