Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.hardware > #3027 > unrolled thread

Which hard drive for raid1 for 24x7 machine

Back to article view | Back to comp.os.linux.hardware

Contents

  Which hard drive for raid1 for 24x7 machine Peter Chant <pete@petezilla.co.uk> - 2016-03-18 18:27 +0000
    Re: Which hard drive for raid1 for 24x7 machine philo <philo@privacy.net> - 2016-03-18 14:03 -0500
      Re: Which hard drive for raid1 for 24x7 machine Peter Chant <pete@petezilla.co.uk> - 2016-03-19 10:05 +0000
        Re: Which hard drive for raid1 for 24x7 machine philo <philo@privacy.net> - 2016-03-19 07:02 -0500
          Re: Which hard drive for raid1 for 24x7 machine Peter Chant <pete@petezilla.co.uk> - 2016-03-19 14:40 +0000
            Re: Which hard drive for raid1 for 24x7 machine philo <philo@privacy.net> - 2016-03-19 12:19 -0500
            Re: Which hard drive for raid1 for 24x7 machine Henrik Carlqvist <Henrik.Carlqvist@deadspam.com> - 2016-03-19 22:19 +0000
              Re: Which hard drive for raid1 for 24x7 machine Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-03-20 03:35 +0000
                Re: Which hard drive for raid1 for 24x7 machine Scott Hemphill <hemphill@hemphills.net> - 2016-03-20 09:31 -0400
              Re: Which hard drive for raid1 for 24x7 machine Dave <noone$$@llondel.org> - 2016-03-24 19:47 -0700
                Re: Which hard drive for raid1 for 24x7 machine philo <philo@privacy.net> - 2016-03-25 05:26 -0500
    Re: Which hard drive for raid1 for 24x7 machine Henrik Carlqvist <Henrik.Carlqvist@deadspam.com> - 2016-03-18 19:39 +0000

#3027 — Which hard drive for raid1 for 24x7 machine

I've got a machine with data stored on two 3TB disks formatted with
btrfs in a raid1 configuration.  Machine is used as a desktop, home
server and does a bit of data crunching periodically with a mariadb
installation.  One disk is showing signs of failure.  I'm not a gamer
but my usage is probably not quite typical  for a home user.

Yes, I am aware of backing up - so we can avoid that discussion.

Given that I have approximately 700GB free on the filesystem I don't
need to go for a disk larger than 3TB, and owing to raid1 I would not
utilise it until I added a further disk to the raid1 array or replace
the other disk with a larger one.

So I'm wondering which one of the three options below is best:

1. Get any cheap 3TB desktop disk.

2. Get a NAS disk intended for 24x7 such as a WD Red.

3. Frighten my wallet and get an enterprise disk at twice the price of
option 1.

Whatever I do it will be paired for now with a desktop or possibly ex
usb enclosure hdd until that disk expires or is redundant.

Thoughts?

Pete

[toc] | [next] | [standalone]

#3028

On 03/18/2016 01:27 PM, Peter Chant wrote:
> I've got a machine with data stored on two 3TB disks formatted with
> btrfs in a raid1 configuration.  Machine is used as a desktop, home
> server and does a bit of data crunching periodically with a mariadb
> installation.  One disk is showing signs of failure.  I'm not a gamer
> but my usage is probably not quite typical  for a home user.
>
> Yes, I am aware of backing up - so we can avoid that discussion.
>
> Given that I have approximately 700GB free on the filesystem I don't
> need to go for a disk larger than 3TB, and owing to raid1 I would not
> utilise it until I added a further disk to the raid1 array or replace
> the other disk with a larger one.
>
> So I'm wondering which one of the three options below is best:
>
> 1. Get any cheap 3TB desktop disk.
>
> 2. Get a NAS disk intended for 24x7 such as a WD Red.
>
> 3. Frighten my wallet and get an enterprise disk at twice the price of
> option 1.
>
> Whatever I do it will be paired for now with a desktop or possibly ex
> usb enclosure hdd until that disk expires or is redundant.
>
> Thoughts?
>
> Pete
>



I think it would be a good idea to get the WD Red  or perhaps an HGST 
Enterprise

Standard desktop drives are so inexpensive that if you bought a drive at 
"twice the price" it still would hardly be a fortune

[toc] | [prev] | [next] | [standalone]

#3030

On 03/18/2016 07:03 PM, philo wrote:

> 
> I think it would be a good idea to get the WD Red  or perhaps an HGST
> Enterprise

Good disk is a WD Red as it turns out, so that really points me to the
choice for the 2nd drive.

> 
> Standard desktop drives are so inexpensive that if you bought a drive at
> "twice the price" it still would hardly be a fortune

Yes and no.  I'm not replacing both disks so I'm tied to the performance
of the lower performer.  It is a thought, but it is a chunk of cash.
Depends on how often I buy bits.

[toc] | [prev] | [next] | [standalone]

#3031

On 03/19/2016 05:05 AM, Peter Chant wrote:
> On 03/18/2016 07:03 PM, philo wrote:
>
>>
>> I think it would be a good idea to get the WD Red  or perhaps an HGST
>> Enterprise
>
> Good disk is a WD Red as it turns out, so that really points me to the
> choice for the 2nd drive.
>
>>
>> Standard desktop drives are so inexpensive that if you bought a drive at
>> "twice the price" it still would hardly be a fortune
>
> Yes and no.  I'm not replacing both disks so I'm tied to the performance
> of the lower performer.  It is a thought, but it is a chunk of cash.
> Depends on how often I buy bits.
>
>


A few days before you posted , I decided that it was time to clone my 
present drive as I'm about to perform an OS upgrade. After doing the 
research , for my system I chose a WD Black.

For your situation, the WD red would be the choice, of course.


Though I have not had problems with the economical "standard" hard 
drives, I really would not recommend them for 24/7 use.

One of my friends, who lost quite a number of drives, never had a 
problem again once I advised him not to keep his machine on 24/7

[toc] | [prev] | [next] | [standalone]

#3032

On 03/19/2016 12:02 PM, philo wrote:
> A few days before you posted , I decided that it was time to clone my
> present drive as I'm about to perform an OS upgrade. After doing the
> research , for my system I chose a WD Black.
> 
> For your situation, the WD red would be the choice, of course.
> 
> 
> Though I have not had problems with the economical "standard" hard
> drives, I really would not recommend them for 24/7 use.
> 
> One of my friends, who lost quite a number of drives, never had a
> problem again once I advised him not to keep his machine on 24/7

I've had failures in the past and this is the first time I've spotted
one on the way out.

I did have a scsi drive once (a long while back) which was rock solid
24/7 for about 7 years.  Only replaced it as its capacity ended up being
ridiculously small.  However, a sample of 1 is hardly meaningful.

[toc] | [prev] | [next] | [standalone]

#3033

On 03/19/2016 09:40 AM, Peter Chant wrote:

> I've had failures in the past and this is the first time I've spotted
> one on the way out.
>

If I catch a read/ write error right away, I can usually save all the 
data. I had one horrible experience though:

I was working on a machine performing a data recovery on one drive and 
somewhere along the line, one of the working drives quit entirely. It 
was not seen in the BIOS.

I broke out in a cold sweat. How was I going to tell the guy that I now 
made things worse?


I pulled the drive out of the machine and went over the circuit board 
with a good magnifying glass. One SMT capacitor was slightly raised on 
one side due to an evidently poor solder joint. It was late at night and 
I had enough sense to put off the repair until the next morning.
I breathed a huge sigh of relief when the drive functioned again.

I then gave the guy a stern lecture about backing everything up to at 
least two separate hard drives and told him I'd never do anything for 
him again unless he did so.








> I did have a scsi drive once (a long while back) which was rock solid
> 24/7 for about 7 years.  Only replaced it as its capacity ended up being
> ridiculously small.  However, a sample of 1 is hardly meaningful.
>

[toc] | [prev] | [next] | [standalone]

#3034

On Sat, 19 Mar 2016 14:40:00 +0000, Peter Chant wrote:
> I did have a scsi drive once (a long while back) which was rock solid
> 24/7 for about 7 years.  Only replaced it as its capacity ended up being
> ridiculously small.  However, a sample of 1 is hardly meaningful.

Back in august 1996 I installed Slackware 3.1 on a 486 with an IDE drive. 
The purpose of that machine was to be an NFS file server and lpd printer 
server for a printer connected to its parallell port. I no longer mount 
its NFS exported drive, but I still use it as my printer server. It has 
been running 24/7 for almost 20 years now and the same IDE drive is still 
running fine. I don't have any UPS so the uptime isn't so impressive, it 
reboots almost every year during the summers when there are thunderstorms.

Even if this is only yet another sample of 1 anectdote I still would say 
that most HDs running 24/7 get retired because of obsolete capacity 
rather than HDD failures.

regards Henrik
-- 
The address in the header is only to prevent spam. My real address is:
hc351(at)poolhem.se Examples of addresses which go to spammers:
root@localhost postmaster@localhost

[toc] | [prev] | [next] | [standalone]

#3035

On 2016-03-19, Henrik Carlqvist <Henrik.Carlqvist@deadspam.com> wrote:

> On Sat, 19 Mar 2016 14:40:00 +0000, Peter Chant wrote:
>
>> I did have a scsi drive once (a long while back) which was rock solid
>> 24/7 for about 7 years.  Only replaced it as its capacity ended up being
>> ridiculously small.  However, a sample of 1 is hardly meaningful.
>
> Back in august 1996 I installed Slackware 3.1 on a 486 with an IDE drive. 
> The purpose of that machine was to be an NFS file server and lpd printer 
> server for a printer connected to its parallell port. I no longer mount 
> its NFS exported drive, but I still use it as my printer server. It has 
> been running 24/7 for almost 20 years now and the same IDE drive is still 
> running fine. I don't have any UPS so the uptime isn't so impressive, it 
> reboots almost every year during the summers when there are thunderstorms.
>
> Even if this is only yet another sample of 1 anectdote I still would say 
> that most HDs running 24/7 get retired because of obsolete capacity 
> rather than HDD failures.

I can add a few more data points - not too many, because the drives last
so damned long.  I'm a believer in 24/7 operation - fewer power surges,
fewer chances for stiction, less thermal cycling...

-- 
/~\  cgibbs@kltpzyxm.invalid (Charlie Gibbs)
\ /  I'm really at ac.dekanfrus if you read it the right way.
 X   Top-posted messages will probably be ignored.  See RFC1855.
/ \  HTML will DEFINITELY be ignored.  Join the ASCII ribbon campaign!

[toc] | [prev] | [next] | [standalone]

#3036

Charlie Gibbs <cgibbs@kltpzyxm.invalid> writes:

[snip]

> I can add a few more data points - not too many, because the drives last
> so damned long.  I'm a believer in 24/7 operation - fewer power surges,
> fewer chances for stiction, less thermal cycling...

I can add another data point.  I am very much in favor of the WD Red
drives.  I burned out a bunch of Seagate drives before I switched to WD
Red in my two 8-drive RAID6 arrays.  I believe the TLER (time limited
error recovery) is important for RAID.

https://www.pugetsystems.com/labs/articles/Western-Digital-Green-vs-Red-Hard-Drives-602/

Scott
-- 
Scott Hemphill	hemphill@alumni.caltech.edu
"This isn't flying.  This is falling, with style."  -- Buzz Lightyear

[toc] | [prev] | [next] | [standalone]

#3037

Henrik Carlqvist wrote:

> On Sat, 19 Mar 2016 14:40:00 +0000, Peter Chant wrote:
>> I did have a scsi drive once (a long while back) which was rock solid
>> 24/7 for about 7 years.  Only replaced it as its capacity ended up being
>> ridiculously small.  However, a sample of 1 is hardly meaningful.
> 
> Back in august 1996 I installed Slackware 3.1 on a 486 with an IDE drive.
> The purpose of that machine was to be an NFS file server and lpd printer
> server for a printer connected to its parallell port. I no longer mount
> its NFS exported drive, but I still use it as my printer server. It has
> been running 24/7 for almost 20 years now and the same IDE drive is still
> running fine. I don't have any UPS so the uptime isn't so impressive, it
> reboots almost every year during the summers when there are thunderstorms.
> 
> Even if this is only yet another sample of 1 anectdote I still would say
> that most HDs running 24/7 get retired because of obsolete capacity
> rather than HDD failures.
> 
A long time ago, possibly back in the MFM days with two cables per drive, a 
friend had a 24/7 machine and one day he came home to the sound of the hard 
drive hunting. He did a frantic backup (only MB, not GB or TB but transfer 
speeds were proportionally lower) and once he'd copied all the data onto 
another machine, he turned that one off. Once stopped, the drive seized and 
never went again, it must have chewed its bearings to paste.

I have a drive giving a couple of SMART errors, I've moved everything 
important off of it but it can stay for now. It was pressed into service as 
the only one I had to hand at the time of building that system so I don't 
know its full service history. In general, when in use my drives all stay up 
24/7 barring powercuts or the occasional exceptional deliberate power down. 
I have probably averaged losing one a year over the past 30 years, thinking 
of the big pile of defunct drives I smashed up a couple of years ago.

-- 
Dave
Too many gadgets, too little time

[toc] | [prev] | [next] | [standalone]

#3038

On 03/24/2016 09:47 PM, Dave wrote:

>
> I have a drive giving a couple of SMART errors, I've moved everything
> important off of it but it can stay for now. It was pressed into service as
> the only one I had to hand at the time of building that system so I don't
> know its full service history. In general, when in use my drives all stay up
> 24/7 barring powercuts or the occasional exceptional deliberate power down.
> I have probably averaged losing one a year over the past 30 years, thinking
> of the big pile of defunct drives I smashed up a couple of years ago.
>



Any drive with a SMART error should be backed up and replaced /at once/

Sure, it might last a few more months but just as likely it could be 
dead tomorrow...why risk it?

BTW; My new WD Black drive arrived earlier this week and I cloned my old 
drive and keep it as a backup. My upgrade from Ubuntu 15.10 to 16.04 
went well. (Two minor, but easily fixed glitches.) Works extremely 
well...and the upgrade kept all my former settings.

When I opened up my machine I had a look at the 2nd HD which still has 
XP ..and noticed it is also a WD Black.


As far as leaving my machine on 24/7 , I'm against the idea...I just 
hibernate and can be back up and running in five seconds

[toc] | [prev] | [next] | [standalone]

#3029

On Fri, 18 Mar 2016 18:27:08 +0000, Peter Chant wrote:
> So I'm wondering which one of the three options below is best:
> 
> 1. Get any cheap 3TB desktop disk.
> 
> 2. Get a NAS disk intended for 24x7 such as a WD Red.
> 
> 3. Frighten my wallet and get an enterprise disk at twice the price of
> option 1.

I would say that the best option for you depends mostly on your need for 
performance. But I would probably go for a cheap desktop drive with 
reasonable speed and stay away for slow archive disks.

As you now have a raid1 configuration and are going to replace 1 of 2 
disks it would also make sense to look at the performance of the drive 
you are going to keep. There is no big point in mixing disks with 
different performance in the same raid configuration.

So my choice would be a cheap drive with long warranty which matches the 
performance of your existing drive in your raid configuration.

regards Henrik
-- 
The address in the header is only to prevent spam. My real address is:
hc351(at)poolhem.se Examples of addresses which go to spammers:
root@localhost postmaster@localhost

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.hardware

csiph-web