Re: ATA / IDE Controller for Blade 1000 / 2000


"Benjamin Gawert" <bgawert@xxxxxx> schrieb im Newsbeitrag
news:4kd2p2FbdogmU1@xxxxxxxxxxxxxxxxx
* Andreas Wacknitz:

Do you know what you are talking about? Three months ago I had a
hard disk failure at the office. I (and thus the company I am working
for) lost most of the work of one day and another day for re-installing
all the needed software on a new drive.

If you need two days for the simple task of reinstalling an office PC then
you're doing something seriously wrong. Where I work we have images of all
computers that are in use. No matter if a simple office PC or a
workstation has a hard drive failure, restore takes usually less than an
hour...

The drive was a one year old SATA drive. So, the company saved approx.
150? for the drive and paid this savings with 2 developer days (approx.
1000?).

Right, but simply because of the lack of a proper backup system...
No, the company has a backup system. But only for the servers. A decent
backup system costs >50,000 EUROS. That's too much for an ordinary
workstation. A RAID system would have been much better but if there isn't
enough money for a decent hard disk, I doubt there would be enough money
for a decent RAID system. So, the company is taking the risk to loose a
day's work if a hard disk breaks, but that seems to be OK for the decision
makers.


I am talking about workstations and not home PCs.
So, from your comments I conclude that your work (or time) is worth just
a few bucks but my work is more expensive.
Furthermore I have decided for myself that my personal data and my
spare time is worth more than the difference between a cheap SATA drive
and a more expensive SCSI drive. Even if the SCSI drive promises just a
little bit more reliability.

You seem to think that SCSI is automatically more reliable which is not.
Even worse, you also seem to think SCSI disks replace a proper backup
system which (sorry) is plain stupid.
No, I don't think so. The company has an appropriate backup system for its
servers. And I have an appropriate backup system for my home use. But a
backup system cannot prevent for any data loss.
Backups are scheduled once a day and not continously (at least not at the
place I used to work).

I expect more reliability of a disk with an MTBF of 1.4 mio hours (spec for
my Seagate SCSI drives) than of a drive with an MTBF of 600,000 hours (spec
for the cheaper SATA drives).


We have several ten thousands of disks at work ((S)ATA, SCSI and now more
and more SAS). SCSI didn't proove itself to be really more reliable than
the (S)ATA disk drives. Of course there are crappy (S)ATA drives out there
but there also are bad series of SCSI drives (i.e. Hitachi DDYS-T). Disks
are unreliable by design, you never know if the drive
You draw conclusions from your experiences; I draw conclusions from my
experiences (two damaged SATA drives; one at home and one at work) and from
the manufactures technical specifications.

lasts forever or fails within the next week. That's why a proper backup is
always necessary, if you don't have a backup you don't have important
data. Period.
Again, I was talking about workstations. I haven't seen a single workstation
with a backup robot, do you?
A server is a very different story.

Andreas


.

Relevant Pages

  • Re: Hard drive stress test
    ... You might look in your hard disk manufacturer's web site for testing ... finding bad sectors on SMD disk drives so they could be mapped out. ... I don't think I ever wrote any code to map out flaws on SCSI drives. ...
    (freebsd-questions)
  • RE: Grub clarification
    ... trouble getting grub to find the drive. ... Some drives and CMOS combinations just do not work. ... be the very first 512 bytes of the disk - loading GRUB on the second ... The SCSI drives or the IDE drives? ...
    (Fedora)
  • Re: ATA / IDE Controller for Blade 1000 / 2000
    ... There is something called imaging software like Acronis Trueimage which is even much cheaper than the DAT72 and which can backup the whole HD content to another storage device or even to a server over the net... ... RAID is a way to increase reliability of a disk subsystem. ... you also seem to think SCSI disks replace a proper backup system which is plain stupid. ... my Seagate SCSI drives) than of a drive with an MTBF of 600,000 hours. ...
    (comp.sys.sun.hardware)
  • Re: Migrate Data from 18Gb SCSI Drive to 36GB SCSI Drive
    ... you can't remove a disk from a VG until all partitions are deleted ... Also, once a mirror is broken, there is only one "live" copy of the data. ... These internal disks are both 18GB scsi drives and I just ...
    (AIX-L)
  • RE: Grub clarification
    ... trouble getting grub to find the drive. ... Some drives and CMOS combinations just do not work. ... be the very first 512 bytes of the disk - loading GRUB on the second ... The SCSI drives or the IDE drives? ...
    (Fedora)