Slightly odd HD problem

[Li'l ol' me]

I've got this Seagate ST340810A (40 gig) that *seems* to work, but behaves
oddly.

I can read and write OK (and so can the drive HO HO!), but reading and
writing is VERY slow (for the drive!) and seems to speed up and slow down.
This is only really noticable when copying large quantities to it, but when
I run Norton Disk Doctor in DOS, it will take 30 mins to check what can be
checked in about 5 seconds on a working disk. Yet there are NO bad sectors
I can see in what it's checked so far.

And the weirdest thing is that I left NDD for 30 minutes, but when I came
back, the 'time lapsed' read only 8 minutes! Watching it in real time, I
could see the clock would only change every few seconds and only one second
would be recorded every few 'real' seconds.

However, it passes windows scandisk except in 'retrieve bad sectors' mode
where it won't even start ie. the progress marker never even appears.

Would a low-level format help? The only m/f diag tools I know that have LLF
utils are from IBM and Maxtor, so I wouldn't know what to use anyway.

 

[Nick]

I've got this Seagate ST340810A (40 gig) that *seems* to work, but behaves
oddly.

I can read and write OK (and so can the drive HO HO!), but reading and
writing is VERY slow (for the drive!) and seems to speed up and slow down.
This is only really noticable when copying large quantities to it, but when
I run Norton Disk Doctor in DOS, it will take 30 mins to check what can be
checked in about 5 seconds on a working disk. Yet there are NO bad sectors
I can see in what it's checked so far.

And the weirdest thing is that I left NDD for 30 minutes, but when I came
back, the 'time lapsed' read only 8 minutes! Watching it in real time, I
could see the clock would only change every few seconds and only one second
would be recorded every few 'real' seconds.

However, it passes windows scandisk except in 'retrieve bad sectors' mode
where it won't even start ie. the progress marker never even appears.

Would a low-level format help? The only m/f diag tools I know that have LLF
utils are from IBM and Maxtor, so I wouldn't know what to use anyway.

There are no such thing as a low level format on IDE drive.

I would advice you to backup all your data on another hard drive since
your Seagate is most likely to die in a short time.

For the time issue you're talking about, it's most likely because NDD
checks in PIO mode hence does a lot of interrupt. Since your hard
drive is close to fail the transfert is time consuming.

Nick

 

[Barry OGrady]

I've got this Seagate ST340810A (40 gig) that *seems* to work, but behaves
oddly.

I can read and write OK (and so can the drive HO HO!), but reading and
writing is VERY slow (for the drive!) and seems to speed up and slow down.
This is only really noticable when copying large quantities to it, but when
I run Norton Disk Doctor in DOS, it will take 30 mins to check what can be
checked in about 5 seconds on a working disk. Yet there are NO bad sectors
I can see in what it's checked so far.

And the weirdest thing is that I left NDD for 30 minutes, but when I came
back, the 'time lapsed' read only 8 minutes! Watching it in real time, I
could see the clock would only change every few seconds and only one second
would be recorded every few 'real' seconds.

However, it passes windows scandisk except in 'retrieve bad sectors' mode
where it won't even start ie. the progress marker never even appears.

Would a low-level format help? The only m/f diag tools I know that have LLF
utils are from IBM and Maxtor, so I wouldn't know what to use anyway.

A low level format is not possible.
Check the power connector. I had a drive that would slow down sometimes, caused
by a poor power connection.


-Barry
========
Web page: http://members.optusnet.com.au/~barry.og
Atheist, radio scanner, LIPD information.
Voicemail/fax number +14136227640

 

[Li'l ol' me]

There are no such thing as a low level format on IDE drive.

I've just started 3 threads simultaneously, and I said in the top one I
don't want to get into a deep discussion of 'real' LLFing. As far as I'm
concerned, erasing by writing lots of zeroes or whatever is LLFing, at least
it's more than a straight format, right?

I would advice you to backup all your data on another hard drive since
your Seagate is most likely to die in a short time.

Luckily there's no data involved, just a drive I got in an untested bundle.

 

[Nick]

I've just started 3 threads simultaneously, and I said in the top one I
don't want to get into a deep discussion of 'real' LLFing. As far as I'm
concerned, erasing by writing lots of zeroes or whatever is LLFing, at least
it's more than a straight format, right?

Erasing the filesystem is quick-formating
Erasing with a lot of zero is formating.
Low level formating includes some more tricks any modern drive would
accept.

Nick

 

[Rod Speed]

I've got this Seagate ST340810A (40 gig)
that *seems* to work, but behaves oddly.

I can read and write OK (and so can the drive HO HO!), but reading
and writing is VERY slow (for the drive!) and seems to speed up and
slow down. This is only really noticable when copying large quantities
to it, but when I run Norton Disk Doctor in DOS, it will take 30 mins
to check what can be checked in about 5 seconds on a working disk.
Yet there are NO bad sectors I can see in what it's checked so far.

Most likely its succeeding on repeated retrys.

See what HDD Health says about the SMART data.

And the weirdest thing is that I left NDD for 30 minutes, but when
I came back, the 'time lapsed' read only 8 minutes! Watching it in
real time, I could see the clock would only change every few seconds
and only one second would be recorded every few 'real' seconds.

Most likely the repeated retrys on partial bads ends up
losing lots of clock ticks, so the time goes bad like that.

However, it passes windows scandisk except in 'retrieve bad sectors' mode
where it won't even start ie. the progress marker never even appears.

What happens if you use the dos mode version of scandisk ?

Would a low-level format help?

Probably not.

The only m/f diag tools I know that have LLF utils are from IBM and Maxtor,

Seagate has those too. Thats what you should be using on a Seagate drive.

 

[Rod Speed]

I'm using Seagate Seatools 1.06.02 which has no LLF/Erase tools.
Correct.

Seagate have a util called DiscWizard,
but this only mentions 'Fast formatting'.
http://www.seagate.com/support/kb/disc/faq/ata_llfmt_what.html

 

[Folkert Rienstra]

There are no such thing as a low level format on IDE drive.

Ofcourse there is. Most drives still support it. What the drive does is another
matter. And yes, there is little point in Low Level Formatting tracks when
you can't change sector size on an IDE drive. You can't erase them, so you
can't actually 'fresh'ly write them as that data is already there.
That isn't to say that you can't actually perform some of the actions that a
'fresh' Low Level Format would perform. Rewriting sector IDs so that LBAs
would be in sequential order (after bad block replacements) is one of them.

 

[Folkert Rienstra]

Erasing the filesystem is quick-formating
Erasing with a lot of zero is formating.

Nope. Formatting actually does very little writing.

 

[Folkert Rienstra]

A low level format is not possible.

Yes it is. It is actually very simple once the tracks have been written.
Servo track writing is(/was) not possible (although IBM claim to use the
drive's own heads to write the tracks), LLF is. SCSI drives support LLF
and they are not that physically different from IDE drives.

Learn what a Low Level Format is.

 

[Folkert Rienstra]

I've got this Seagate ST340810A (40 gig) that *seems* to work, but behaves
oddly.

I can read and write OK (and so can the drive HO HO!), but reading and
writing is VERY slow (for the drive!) and seems to speed up and slow down.
This is only really noticable when copying large quantities to it, but when
I run Norton Disk Doctor in DOS, it will take 30 mins to check what can be
checked in about 5 seconds on a working disk. Yet there are NO bad sectors
I can see in what it's checked so far.

And the weirdest thing is that I left NDD for 30 minutes, but when I came
back, the 'time lapsed' read only 8 minutes! Watching it in real time, I
could see the clock would only change every few seconds and only one second
would be recorded every few 'real' seconds.

I think that interrupts are off when I/O is done. When the drive does
repeated retries, the IO doesn't finish and the clock isn't updated in time.
I believe this also why you see a system freeze when it hits bad sectors.

However, it passes windows scandisk except in 'retrieve bad sectors' mode

What windows is that?

where it won't even start ie. the progress marker never even appears.

Would a low-level format help?

Probably. Any write to the bad sectors would, but that is not exactly
an easy feat unless it would be for instance the restoration of a backup.

The only m/f diag tools I know that have LLF
utils are from IBM and Maxtor, so I wouldn't know what to use anyway.

Any ute that writes zeroes would be fine.
Bart's Disktool is such a ute and it can rigorously test your drive also.
www.nu2.nu/utilities

 

[Li'l ol' me]

I think that interrupts are off when I/O is done. When the drive does
repeated retries, the IO doesn't finish and the clock isn't updated in time.
I believe this also why you see a system freeze when it hits bad sectors.
mode

What windows is that?


Probably. Any write to the bad sectors would, but that is not exactly
an easy feat unless it would be for instance the restoration of a backup.


Any ute that writes zeroes would be fine.
Bart's Disktool is such a ute and it can rigorously test your drive also.
www.nu2.nu/utilities

Thanks, but that's a dead link when I tried.

 

[Folkert Rienstra]

Which Windows?

http://www.nu2.nu/utils/


Thanks, but that's a dead link when I tried.

It was down. It is now up again.

 

[Nick]

Nope.

I bet LLF the Track 0 would be a very, very bad idea.

Nick

 

[Folkert Rienstra]

I bet

You shouldn't be betting with your level of knowledge.

LLF the Track 0 would be a very, very bad idea.

You loose. The selfdestructing harddrive huh?
Are you really that dense that you think that the drive would happily
destroy it's own reserved area? And for what? The reserved area
contains structures that need to stay where the firmware expects them.
There is absolutely no point in LLFing that area.

 

[Fred]

Utterly clueless. A LLF doesn't operate on the reserved area, only the user
area. The reserved area on my IBM DMVS09 is 21*5 tracks=110 tracks.
Nothing like a single track. LBA 0 is at Physical CHS 0x16/0x0/0x0.

Utterly clueless, LBA 0 is at Physical CHS 0x16/0x0/0x1

 

[Folkert Rienstra]

Or more specifically, the reserved area ends at physical 0x15/0x4/0x17D
(or 21/4/381 in decimal).

Utterly clueless, LBA 0 is at Physical CHS 0x16/0x0/0x1

Roddles, Roddles. There was a reason why I said "Physical" CHS.
No wonder you didn't dare use your own name.

 

[Folkert Rienstra]

Nope, physical sectors start with 0 unlike with BIOS logical CHS.
LBA 0 obviously is CHS 0/0/1 in BIOS logical CHS.

I was speaking of Track 0, which is not LBA 0.

So? No one said that it was.

The track 0 is what the hard drive reads when it's seeking.

That's a load of crap. Seeks would take forever if it did that.

And anyway speaking in CHS is obsolete.

No kidding. Physical CHS is very much alive, internal to the harddrive.
How else would the firmware know where a specific LBA is located?
Sector however may be obsolete now that NoID sectors are the norm.
A sector address is now an offset into the start of a track.

 

[Fred]

Nope,

Yep, utterly clueless.

physical sectors start with 0

Wrong. As always.

unlike with BIOS logical CHS.

Wrong. As always.

LBA 0 obviously is CHS 0/0/1 in BIOS logical CHS.

And physical CHS too.

 

[Fred]

Or more specifically, the reserved area ends at physical 0x15/0x4/0x17D
(or 21/4/381 in decimal).

Roddles, Roddles.

****nut, ****nut, ****nut.

There was a reason why I said "Physical" CHS.

Pity you're still utterly clueless.

Even you should be able to bullshit your way out of your
utter cluelessness better than that pathetic effort, ****nut.