more about bad sectors

[Irwin]

I may want to use ghost with the following switches to copy my drive
with bad sectors, but I don't understand the difference. Can someone
please explain the difference? Which one is more appropriate, or should
I combine them and use both? It sorta sounds like the BFC switch
examines the target, but having read other newsgroup messages that
doesn't make sense.


-BFC
Handles bad FAT clusters when writing to disk.
If this switch is set, and the target partition is FAT, Ghost checks
for and skips up to 500 bad sectors, and creates a file that lists the
bad sectors. This option may substantially slow Ghost operations.
Without this switch, Ghost aborts when Ghost encounters a bad sector.

-FRO
Forces Ghost to continue cloning even when the source disk
contains bad blocks.

Thanks,
Irwin

 

[DEMAINE Benoit-Pierre]

IMHO, BFC seems to scan BB on target drive, and FRO on source.

Scanning target sould not make sense for new bought drives which aimed to be fault
free. It does make sens to me, since 'any' drive could be faulty and have BB. I
could crash 3 new drives in 48h ... so BB scaning on target drive does make sens to me.

If your src drive has BB, you do want FRO.

BFC is usefull to check quality of target, non sero count should be interpreted
carefully:

non wero BFC count means the target is faulty, AND NOT RELIABLE FOR USE, and should
be replaced ASAP. Still, when you are about to introduce a 3rd disk (hopefully fault
less), you do not want to read again the oldest faulty source, because precedent
readings may have alter worse the datas: you may want to use the 2nd drive as source
for the 3rd one, since it will have as many errors (if any).

Trying to read twice the most faulty drive may result in severe data corruption.

I susually make multiple copy, and use dedicated tools to compare content of files.

To prevent badblocks, all my computer run SMART monitoring tools. <8y drives perform
self BB relocation: if the firmware thinks a sector is faulty, it will automatically
remap it somewhere else, and move data all by itself. You start having data loss
when you have more physical BB than room in the relocation space; until that space
is full, you do have BB, but no data loss. SMART tools are a software interface to
tell you how many sectors have been moved. Not reading those are like looking at the
sword upon your head, and waiting for it to fall on you: irresponsible.

 

[Irwin]

Demaine, thanks for the advice. What smart tools are you currently
using?

IMF

 

[DEMAINE Benoit-Pierre]

Demaine, thanks for the advice. What smart tools are you currently
using?

IMF

The standard package available for my distribution.

For Gentoo 2005.1 ~x86, it seems to be:

* sys-apps/smartmontools
Latest version available: 5.33
Latest version installed: 5.33
Size of downloaded files: 497 kB
Homepage: http://smartmontools.sourceforge.net/
Description: control and monitor storage systems using the Self-Monitoring,
Analysis and Reporting Technology System (S.M.A.R.T.)
License: GPL-2

Debian also have only one package available. Comes with smartctl, and a daemon.
Daemon fills syslog. It is your responsability to use a filter and email yourself
revelant parts. I dont use log filters because I have bits of syslog all the time on
screen, I do myself my live monitoring, visually, real time.

I found soem GUI for Ms-windows which I did not like at all, because they lacked
system integration, not installed as service even into NT5, and crapy GUI. Further
more, there was no free toll ever developped, only trials time limitted.

As if poor people unable to afford for softwares had to suffer disk crashes. SMART
should always have been integrated as non removable driver of all systems, and
produce red alert to any user logged in. It is so easy: make an ACPI call, read
SMART data, open a pop up !!! and loads of user loose data, files, and disks because
nobody ever think to install SMART tools.

We would not even need RAID if admins did use SMART properly ! Maybe thats a bit
extremistic view ... but think about what I mean.

 

[Rod Speed]

The standard package available for my distribution.

For Gentoo 2005.1 ~x86, it seems to be:

* sys-apps/smartmontools
Latest version available: 5.33
Latest version installed: 5.33
Size of downloaded files: 497 kB
Homepage: http://smartmontools.sourceforge.net/
Description: control and monitor storage systems using the
Self-Monitoring, Analysis and Reporting Technology System (S.M.A.R.T.)
License: GPL-2

Debian also have only one package available. Comes with smartctl, and
a daemon. Daemon fills syslog. It is your responsability to use a
filter and email yourself revelant parts. I dont use log filters
because I have bits of syslog all the time on screen, I do myself my
live monitoring, visually, real time.

I found soem GUI for Ms-windows which I did not like at all, because
they lacked system integration, not installed as service even into
NT5, and crapy GUI. Further more, there was no free toll ever
developped, only trials time limitted.

As if poor people unable to afford for softwares had to suffer disk
crashes. SMART should always have been integrated as non removable
driver of all systems, and produce red alert to any user logged in.
It is so easy: make an ACPI call, read SMART data, open a pop up !!!
and loads of user loose data, files, and disks because nobody ever
think to install SMART tools.

We would not even need RAID if admins did use SMART properly !

RAID1 and 5 handle drives that fail without
any prior warning in the SMART data. That
is a substantial percentage of drive failures.

Maybe thats a bit extremistic view ...

Just plain wrong, actually.

but think about what I mean.

You've just got it completely wrong.

 

[Rod Speed]

I may want to use ghost with the following switches to copy
my drive with bad sectors, but I don't understand the difference.
Can someone please explain the difference? Which one is more
appropriate, or should I combine them and use both? It sorta
sounds like the BFC switch examines the target,

Yes, it handles bads on the target, not the source.

but having read other newsgroup messages that doesn't make sense.

Then they are mangling the story.

-BFC
Handles bad FAT clusters when writing to disk.
If this switch is set, and the target partition is FAT, Ghost checks
for and skips up to 500 bad sectors, and creates a file that lists the
bad sectors. This option may substantially slow Ghost operations.
Without this switch, Ghost aborts when Ghost encounters a bad sector.

Its arguable how useful that is, makes more sense to not use the drive.