Fixing not validated HDs
Samy Merchi
Every week or so when some application manages to crash the system
while the HD's writing, my hard drive goes not validated.
So far, I keep fixing this with applying first Disksalv's 'validate'
procedure, but this leaves an odd problem: All files are accessible,
but their icons don't show up. So I use ReOrg after this, and I get
everything back to normal.
Now, my question is -- there -has- to be a program that does both of
these tasks? How do you all fix your HD when it goes not validated?
And is there any way to do this faster than the 45 minutes or so total
that running both these two programs over my 120M partition takes?
-----------------------------------------------------------------
- Samy Merchi 1998 sam...@utu.fi http://www.utu.fi/~samerc
"Reignfire is and has ALWAYS been Roberto Da Costa."
Roberto Da Costa aka Sunspot aka Reignfire, X-Force #43
"JFM absolves Roberto of being Reignfire, and heads'll roll."
Samy Merchi, April 18th 1998, anticipating X-Force #80
Michael Ravn
>Every week or so when some application manages to crash the system
>while the HD's writing, my hard drive goes not validated.
>So far, I keep fixing this with applying first Disksalv's 'validate'
>procedure, but this leaves an odd problem: All files are accessible,
>but their icons don't show up. So I use ReOrg after this, and I get
>everything back to normal.
>Now, my question is -- there -has- to be a program that does both of
>these tasks? How do you all fix your HD when it goes not validated?
>And is there any way to do this faster than the 45 minutes or so total
>that running both these two programs over my 120M partition takes?
I use Amiback Tools. This works normally.
--
Michael Ravn
http://www.image.dk/~mravn
Amiga 1200/Blizzard 1230-IV/18 meg.
Remove REMOVE when replying by E-mail
Paul Hill
On 19-May-98 20:53:40, Samy Merchi talked about Fixing not validated HDs:
>Every week or so when some application manages to crash the system
>while the HD's writing, my hard drive goes not validated.
When ever I get an invalidated hard disk the OS always re-validates (well,
99% of the time) itself. Only rarely do I use disksalv.
Your hard drive shouldn't really be invalidating itself this much. Do you
run some dodgy programs?
>So far, I keep fixing this with applying first Disksalv's 'validate'
>procedure, but this leaves an odd problem: All files are accessible,
>but their icons don't show up. So I use ReOrg after this, and I get
>everything back to normal.
Never seen this problem of icons not showing up. Care to give some more
info?
>Now, my question is -- there -has- to be a program that does both of
>these tasks? How do you all fix your HD when it goes not validated?
>And is there any way to do this faster than the 45 minutes or so total
>that running both these two programs over my 120M partition takes?
I agree DiskSalv is *far* too slow. QuarterBack & AmiBack tools come to
mind as an alternative.
- Paul
Samy Merchi
Paul Hill (pa...@lagernet.clara.co.uk) typed:
> On 19-May-98 20:53:40, Samy Merchi talked about Fixing not validated HDs:
> >Every week or so when some application manages to crash the system
> >while the HD's writing, my hard drive goes not validated.
>
> When ever I get an invalidated hard disk the OS always re-validates (well,
> 99% of the time) itself. Only rarely do I use disksalv.
Well, when I crash and reboot to find the hd invalid, it boots
normally (though way slower than normal) and then when the boot-up is
finished, the disk drive light stays on for several minutes
(presumably validating the disk) until I get a suspend/reboot
requester that task DH0 has crashed. And this happens every time that
I boot with an invalid disk.
> Your hard drive shouldn't really be invalidating itself this much. Do you
> run some dodgy programs?
Just programs that crash with the normal degree of Amiga
<sarcasm>stability</sarcasm>. I'm probably going to have to trade into
a PC soon and run Linux -- the Amiga is way way too unstable for my
purposes, i.e. running a server.
> >So far, I keep fixing this with applying first Disksalv's 'validate'
> >procedure, but this leaves an odd problem: All files are accessible,
> >but their icons don't show up. So I use ReOrg after this, and I get
> >everything back to normal.
>
> Never seen this problem of icons not showing up. Care to give some more
> info?
Well, it's rather weird. After I do the 'validate' on Disksalv, if I
open a window of the HD, it's completely empty. Not having an invalid
disk right now, I can't remember whether the window's titlebar went to
0% full of the disk, or whether it showed the right amount. But in any
case, all programs are accessible as normal via my Startmenu, or if I
type their address accurately in a shell. I also can't recall if this
also extended to shell -- whether a 'dir' would give an empty
listing -- but I seem to recall this was the case.
> >Now, my question is -- there -has- to be a program that does both of
> >these tasks? How do you all fix your HD when it goes not validated?
> >And is there any way to do this faster than the 45 minutes or so total
> >that running both these two programs over my 120M partition takes?
>
> I agree DiskSalv is *far* too slow. QuarterBack & AmiBack tools come to
> mind as an alternative.
I'll look into them. Thanks.
Sinisa Lolic
> On 19-May-98 21:53:40, Samy Merchi said: <
SM> Every week or so when some application manages to crash the system
SM> while the HD's writing, my hard drive goes not validated.
How about using DiskSafe?
Vegi XXX
Tim Corringham
Sinisa Lolic (ve...@usa.net) wrote:
You may be interested to know that there is a new improved version
of AFS about to be released. It has been renamed (back) to
PFS2 (professional file system 2). You'll never have unvalidated
partitions with PFS2.
Tim
--
======================================================================
Tim Corringham e-mail: t...@ramjam.u-net.com
Ramjam Consultants Ltd tel: +44 (0)118 946 5940
Reading, England fax: +44 (0)118 946 5941
Samy Merchi
=?iso-8859-1?Q?J=F6rg?= Strohmayer (nos...@gmx.de) typed:
>
> Also a bug of DirCacheFFS (especially with FFS < 43.19), the DirCache
> blocks are destroyed/empty (list, dir, ... gives you empty dirs or does
> not show all files) but the dir- and file-header blocks are still ok,
> so you can access the files/dirs with their names directly.
>
> Using ReOrg will rebiuld the DirCache blocks, after using ReOrg dir,
> list, ... will work again.
>
> DON'T USE DIRCACHE FFS WITH FFS < 43.19, better don't use it at all,
> validating is _much_ faster too with Intl.-FFS.
Aha! Thanks a lot! Can the filesystem be converted from Dircache FFS
to Intl FFS without reformatting the disks?
A.P. Suggitt
Samy Merchi (sam...@utu.fi) wrote:
> =?iso-8859-1?Q?J=F6rg?= Strohmayer (nos...@gmx.de) typed:
> > DON'T USE DIRCACHE FFS WITH FFS < 43.19, better don't use it at all,
> > validating is _much_ faster too with Intl.-FFS.
>
> Aha! Thanks a lot! Can the filesystem be converted from Dircache FFS
> to Intl FFS without reformatting the disks?
>
See ReOrg_English.guide under "FileSystem Conversion" or equivalent.
peter
--
Peter Suggitt <pe...@apsayton.demon.co.uk> | Suggitt's Ices
Warped A4000 -- "Multimedia by design" | Great Ayton
| North Yorkshire
Scumvermin Rule OK!
Hans Guijt
Paul Hill (pa...@lagernet.clara.co.uk) wrote:
>When ever I get an invalidated hard disk the OS always re-validates (well,
>99% of the time) itself. Only rarely do I use disksalv.
This is an important point. I have a feeling many people don't know this, but
you can simply wait out the validation time and 99% of the time the OS makes
everything alright on its own. This _doesn't_ happen if you cancel the
requester.
>>Now, my question is -- there -has- to be a program that does both of
>>these tasks? How do you all fix your HD when it goes not validated?
Wait. Click 'retry' once the HD led stops flickering. Everything will be
alright, in just a few minutes.
It's almost the same as in Windows 95 really. If you stop the machine while
writing to the HD, you get to sit through the revalidation of the HD.
Hans
Kamil Iskra
A.P. Suggitt wrote:
> > Aha! Thanks a lot! Can the filesystem be converted from Dircache FFS
> > to Intl FFS without reformatting the disks?
> See ReOrg_English.guide under "FileSystem Conversion" or equivalent.
There is a simpler way, if you have a disk editor, like DPU: go to the
bootblock of the partition that makes trouble and change the first 4
bytes from 444f5305 to 444f5303. Let it revalidate, and this time it
won't crash, and you will get a FFS Intl filesystem.
--
/ Kamil Iskra AmigaOS Linux/i386 Linux/m68k \
| GeekGadgets GCC maintainer UNIX system administrator |
| is...@student.uci.agh.edu.pl kis...@ernie.icslab.agh.edu.pl |
\ ka...@dwd.interkom.pl http://student.uci.agh.edu.pl/~iskra /
Larry McGahey
> Samy Merchi <sam...@utu.fi> wrote:
>
> > Well, when I crash and reboot to find the hd invalid, it boots
> > normally (though way slower than normal) and then when the boot-up is
> > finished, the disk drive light stays on for several minutes
> > (presumably validating the disk) until I get a suspend/reboot
> > requester that task DH0 has crashed. And this happens every time that
> > I boot with an invalid disk.
>
Get Disksafe from Aminet and avoid almost all these problems.
Larry
Paul Hill
On 21-May-98 09:16:24, Hans Guijt talked about Re: Fixing not validated HDs:
>Paul Hill (pa...@lagernet.clara.co.uk) wrote:
>>When ever I get an invalidated hard disk the OS always re-validates (well,
>>99% of the time) itself. Only rarely do I use disksalv.
>This is an important point. I have a feeling many people don't know this, but
>you can simply wait out the validation time and 99% of the time the OS makes
>everything alright on its own. This _doesn't_ happen if you cancel the
>requester.
What requester? The 'volume is not validated' requester? This only appears
when a program trys to write to an invalidated disk. Canceling this
requester should only send an error code to the program in question.
>>>Now, my question is -- there -has- to be a program that does both of
>>>these tasks? How do you all fix your HD when it goes not validated?
>Wait. Click 'retry' once the HD led stops flickering. Everything will be
>alright, in just a few minutes.
>It's almost the same as in Windows 95 really. If you stop the machine while
>writing to the HD, you get to sit through the revalidation of the HD.
This only happens with the later versions of Windoze 95. This can be turned
off too. The difference with Win95 though, is that it stops the machine from
booting, it would be nice if AmigaOS did this too.
- Paul
Hannu Nevalainen
In a message dated 22-Maj-98 11:15:08
Kamil Iskra <is...@student.uci.agh.edu.pl> wrote:
>A.P. Suggitt wrote:
>> > Aha! Thanks a lot! Can the filesystem be converted from Dircache FFS
>> > to Intl FFS without reformatting the disks?
>> See ReOrg_English.guide under "FileSystem Conversion" or equivalent.
>There is a simpler way, if you have a disk editor, like DPU: go to the
>bootblock of the partition that makes trouble and change the first 4
>bytes from 444f5305 to 444f5303. Let it revalidate, and this time it
>won't crash, and you will get a FFS Intl filesystem.
Isn't the directory cache blocks stored on disk?
What happens with them if they are?
I have really hard to believe that you get a 100% functional FFS-INTL
partition the way you describe.
A hack is ok for temporary solutions, but you better avoid it for long time
use.
.---- Remove before use /Hannu E K Nevalainen/
___/__
mailto:he...@warez.it.kth.se,
http://www.ztrip.it.kth.se/~henk/
Personal rc5-stats: 4,713 blocks of keys, i.e. 1,265,136,304+ kkeys processed.
--
AmigaOS - The Sharkey OS? (tm)
Sinisa Lolic
> On 21-May-98 00:42:05, Tim Corringham said: <
TC> You may be interested to know that there is a new improved version
TC> of AFS about to be released. It has been renamed (back) to
TC> PFS2 (professional file system 2). You'll never have unvalidated
TC> partitions with PFS2.
I hope there`ll be some repair/optimization tools for it this time...
Vegi XXX
Sinisa Lolic
> On 21-May-98 14:50:06, Samy Merchi said: <
SM> Aha! Thanks a lot! Can the filesystem be converted from Dircache FFS
SM> to Intl FFS without reformatting the disks?
Yes, easily, with ReOrg...
Vegi XXX
Joel Alvim
On 21 May 98 10:16:24 +0100, Hans Guijt wrote some wise words on comp.sys.amiga.applications about Re: Fixing not validated HDs:
> This is an important point. I have a feeling many people don't know this, but
> you can simply wait out the validation time and 99% of the time the OS makes
> everything alright on its own. This _doesn't_ happen if you cancel the
> requester.
I had this problem once. My HD had an error in one block, and when
the drive got invalidated, it tried to validate as ususally. But when
it came across that damaged block, the drive just locked and the
validation jammed. This means, I couldn't boot the drive.
A very complicated process with several pieces of software followed
to get back my data, and I managed to do it :)
Bye!
--
This is
/\/\ind / Amiga User
/ \/\/alker Module Composer
signing off............
WWW: http://www.terravista.pt/enseada/1356/
Joel Alvim
On 20 May 98 16:57:09 +0100, Sinisa Lolic wrote some wise words on comp.sys.amiga.applications about Re: Fixing not validated HDs:
> SM> Every week or so when some application manages to crash the system
> SM> while the HD's writing, my hard drive goes not validated.
>
> How about using DiskSafe?
Disksafe improves things, but my HD still gets invalidated a lot.
Prime reasons for this are computer crashes while writing to disk
(Voyager, IBrowse do it a lot). Then, if there's a power spike during
a write process, disksafe fails either.
Seems like Disksafe works fine when you cause a reset yourself. But
when it's one of those hard locks, it fails hopelessly... ;(
Hans Guijt
Kamil Iskra (is...@student.uci.agh.edu.pl) wrote:
>There is a simpler way, if you have a disk editor, like DPU: go to the
>bootblock of the partition that makes trouble and change the first 4
>bytes from 444f5305 to 444f5303. Let it revalidate, and this time it
>won't crash, and you will get a FFS Intl filesystem.
<ungl!> That must be the most evil data-destroying hack I've seen posted in a
long time...
If you do this MAKE A BACKUP FIRST. Who knows what will happen...
Hans
Hans Guijt
Paul Hill (pa...@lagernet.clara.co.uk) wrote:
>>This is an important point. I have a feeling many people don't know this,
>>but you can simply wait out the validation time and 99% of the time the OS
>>makes everything alright on its own. This _doesn't_ happen if you cancel the
>>requester.
>What requester? The 'volume is not validated' requester? This only appears
>when a program trys to write to an invalidated disk. Canceling this
>requester should only send an error code to the program in question.
Uhm, good point. You only get the requester if you write to the HD. My
startup-sequence does that (for at least some volumes), and I usually choose
to wait. For the other volumes - well, I only find out if it is revalidating
if I note the HD led flickering.
>This only happens with the later versions of Windoze 95. This can be turned
>off too. The difference with Win95 though, is that it stops the machine from
>booting, it would be nice if AmigaOS did this too.
Make sure your startup-sequence writes to the HD at an early stage. In a pinch
something lame like
echo "Revalidate, you slimy bastard!" > hd0:revalidate
echo "Revalidate, you slimy bastard!" > hd1:revalidate
echo "Revalidate, you slimy bastard!" > hd2:revalidate
echo "Revalidate, you slimy bastard!" > hd3:revalidate
...etc
at the beginning of your startup-sequence would do this just fine.
Hans
Paul Hill
On 23-May-98 07:59:12, Hans Guijt talked about Re: Fixing not validated HDs:
Nice, but...This would appear on my 'other' non-cybergraphics monitor which
is normally either switched off, plugged into an XT or my old A2000.....
Really, the validation process doesn't bother me at all. It only happens
rarely, and I have a fast machine with a fast SCSI controler so I hardley
notice.
- Paul
Si
> Hans Guijt <hgu...@inter.nl.net> wrote:
>
> > Kamil Iskra (is...@student.uci.agh.edu.pl) wrote:
> > >There is a simpler way, if you have a disk editor, like DPU: go to the
> > >bootblock of the partition that makes trouble and change the first 4
> > >bytes from 444f5305 to 444f5303. Let it revalidate, and this time it
> > >won't crash, and you will get a FFS Intl filesystem.
> >
> > <ungl!> That must be the most evil data-destroying hack I've seen posted in a
> > long time...
>
> other change will destroy the partition).
>
> > If you do this MAKE A BACKUP FIRST. Who knows what will happen...
>
> same as DirCacheFFS, only without DirCache blocks) and after the next
> validate all (now unused) DirCache blocks will be marked as free in the
> bitmap.
Sorry if this is a stupid question: When I first installed my HD I
didn't know the MaxTransfer value had to be changed. Consequently,
the largest partition on my drive uses the Old File System (the only
one that would work at the time). I don't have sufficient space to
back it up and reformat it, so could I use this hack?
Cheers,
--
Si <s...@2fx.u-net.com>
Kirk Strauser
On 23-May-98 09:47:29, Paul Hill wrote:
>On 23-May-98 07:59:12, Hans Guijt said:
>>echo "Revalidate, you slimy bastard!" > hd0:revalidate
>>...etc
>Nice, but...This would appear on my 'other' non-cybergraphics monitor
>which is normally either switched off, plugged into an XT or my old
>A2000.....
Actually, since all of the output is redirected, you wouldn't ever see
it. I'm like you, though; validation on my 2GB partitions doesn't take
more than 30sec. or so normally. It's not a big bother to me.
Kirk Strauser Member // http://www.pcis.net/kstrauser/
kstr...@pcis.net Team AMIGA \X/ http://csc.smsu.edu/~strauser/RA.html
"Classic" 3000/060/50/604e/200/272MB/Picasso II/IOExtender/Major Attitude
Kirk Strauser
>Sorry if this is a stupid question: When I first installed my HD I
>didn't know the MaxTransfer value had to be changed. Consequently,
>the largest partition on my drive uses the Old File System.
OK, I'm curious. Care to go into detail? :)
>I don't have sufficient space to back it up and reformat it, so could I
>use this hack?
Yes. However, it'd almost 100% guaranteed destroy all of your data. You
could use it, though. ;)
Seriously, get ReOrg (which I think should manditorily come with OS 4+).
It has the option to convert OFS -> FFS, and your drive will be all nice
and happy afterward, too.
Paul Hill
On 24-May-98 13:05:17, Si talked about Re: Fixing not validated HDs:
>> > >bootblock of the partition that makes trouble and change the first 4
>> > >bytes from 444f5305 to 444f5303. Let it revalidate, and this time it
>> ^ *reboot* !
>> > >won't crash, and you will get a FFS Intl filesystem.
>> >
>> > <ungl!> That must be the most evil data-destroying hack I've seen posted
>> > in a long time...
>>
>> No, it works without problems (but only DirCacheFFS -> Intl. FFS, any
>> other change will destroy the partition).
>>
>> > If you do this MAKE A BACKUP FIRST. Who knows what will happen...
>>
>> After a reboot the filesystem will be FFS Intl. (which is exactly the
>> same as DirCacheFFS, only without DirCache blocks) and after the next
>> validate all (now unused) DirCache blocks will be marked as free in the
>> bitmap.
>Sorry if this is a stupid question: When I first installed my HD I
>didn't know the MaxTransfer value had to be changed. Consequently,
>the largest partition on my drive uses the Old File System (the only
>one that would work at the time). I don't have sufficient space to
>back it up and reformat it, so could I use this hack?
NO, NO, NO, NO, NO!
Whatever you do don't try this. OFS and FFS are quiet different. You
CAN use RE-Org to do this however, as a couple of other people have
mentioned.
- Paul
Michael Wilkins
On 20 May 98 16:57:09 +0100, "Sinisa Lolic" <ve...@usa.net> wrote:
> > On 19-May-98 21:53:40, Samy Merchi said: <
>
> SM> Every week or so when some application manages to crash
the system
> SM> while the HD's writing, my hard drive goes not validated.
>
> How about using DiskSafe?
>
> Vegi XXX
>
Hi,
This may not fix everything but it may help. Get "WaitValidate"
from Aminet.
Michael Wilkins
Amiga A4000T 112mb fast/CyberMkII 50Mhz/060/RetinaZ3/Atapi x12 CdRom
Michael Wilkins, Kyoto. JAPAN
emailto:mich...@mbox.kyoto-inet.or.jp
Quantum & Seagate HDs'4.9 Gig.Vlab.HP5l.Supra33.6.IomegaZip.
MicrotekV600 scanner. HewlettPackard HP5l printer.
http://web.kyoto-inet.or.jp/people/michaelb
Hans Guijt
Paul Hill (pa...@lagernet.clara.co.uk) wrote:
>>echo "Revalidate, you slimy bastard!" > hd0:revalidate
>Nice, but...This would appear on my 'other' non-cybergraphics monitor which
>is normally either switched off, plugged into an XT or my old A2000.....
Well, you either get the requester or you don't - it's your own choice ;-)
Hans
Sinisa Lolic
> On 22-May-98 16:01:15, Joel Alvim said: <
>> SM> Every week or so when some application manages to crash the system
>> SM> while the HD's writing, my hard drive goes not validated.
>> How about using DiskSafe?
JA> Disksafe improves things, but my HD still gets invalidated a lot.
I really cannot say when I saw this happen last time, but it was quite a some
time ago...
JA> Prime reasons for this are computer crashes while writing to disk
JA> (Voyager, IBrowse do it a lot). Then, if there's a power spike during
JA> a write process, disksafe fails either.
Well, have you tried to increase a stack value in an .icon file of those
programs? It is known that MUI programs need at least 8kb of stack.
StackAttack program could also help.
JA> Seems like Disksafe works fine when you cause a reset yourself. But
JA> when it's one of those hard locks, it fails hopelessly... ;(
Well, it might be that you have a MaxTransfer problem. Try to decrease this
value in HDToolBox or with CHUNKSIZE argument of DiskSafe.
Vegi XXX