Block ID errors, how do I avoid them?

[-Acid-]

I've successfully setup dozens of CF cards using both SFS and PFS3-AIO so I know how to do it. Max transfer and masks are no problem I know what settings should be used. I usually use idefix97 to enable the partitions beyond 4GB (as I have a buffered IDE interface for my DVD drive) and I use Classic WB as well.

For over a year i've had a 160gb drive in my A1200 with 128gb of it partitioned and it has worked without problems the whole time. I would estimate there to be 30-35gb on the drive but now all of a sudden a block id error problem has cropped up. I get the same problem with load of other drives I've got lying around (mostly 40gb drives).

No matter if I use SFS or AIO, I make the partitions of 2gb then 5gb, 5gb, whatever etc. The first thing I do is attempt to write my ADF collection to one of the 5gb partitions, it gets about half way through then block id errors. This happens if I do it in Winuae or on real hardware.

So am I missing something here? Is there actually a limit on the drive size that can be used? I've also got a 20gb drive that is as old as the hills which never gives me any problems (SFS formatted). Were older drives made differently from later ATA drives which could be affecting things? Would I have better luck using Sata drives with and IDE to SATA converter? does 2.5" and 3.5" form factors have any affect on this problem? Would SCSI drives help? I have the SCSI module for the Cyberstorm in my A4000 but no SCSI drives.

This is sending me round the bend, 4gb CF cards work fine when I set them up, why are large drives being such a problem?

[thomas]

If it happens in WinUAE with many different drives, then it can hardly be a hardware problem.

In order to definitely rule out the hardware you should make a 10 GB HDF and repeat your procedure with this. If it still happens, please post your WinUAE config and a step-by-step guide how to reproduce the errors.

[Daedalus]

Getting halfway through the first 5GB partition sounds suspiciously close to hitting the 4GB mark on the disc structure itself (2GB 1st partition + 2GB, 1st half of 2nd partition). Are you sure scsi.device is being patched correctly by IDEFix on startup? Open a shell and version scsi.device. Try running Check4GB on the system and see if it gives you warnings about any partitions. Both the scsi.device and the filesystem have to be >4GB-capable, so maybe IDEFix isn't installing its patch for some reason. Have you disabled the reboot after IDEFix and Setpatch have run by any chance?

[Toni Wilen]

As thomas already said, more details needed. Something is missing.. Do you use some (buggy?) scsi.device patches?

PFS3AIO shouldn't mount the partition if it crosses 4G (or ~8G if A600/A1200/A4000 IDE) border without required 64-bit support.

You could try backing up your 120gb drive on a PC using WinImage and writing it back to another 120gb drive.

[-Acid-]

Running Check4GB says each partition is running SFS 1.279 and all have ok in the check column. Scsi.device version is 119.16. IDEfix does indeed reboot immediately after first boot to mount the other partitions.

My partitions and how much they have written on them as it stands is

DH0: 1.5Gb 140Mb used
DH1: 63Gb 5.8Gb used
DH2: 63Gb 25.8Gb used

[fitzsteve]

Are you sure drive geometry is correct, my maths says you can't fit your above listed partitions inside 120gb which in fact would be something smaller like 113gb

[fitzsteve]

Sorry, the post above confused me. I see you're running a 160gb drive...

[-Acid-]

Originally Posted by fitzsteve

Sorry, the post above confused me. I see you're running a 160gb drive...

Yes Steve it's a 160Gb drive but I have only partitioned 128Gb of it. I'm sure I read something about that being the maximum drive or volume size for SFS and thought that was more than enough for an Amiga any way.

[-Acid-]

Originally Posted by Toni Wilen

As thomas already said, more details needed. Something is missing.. Do you use some (buggy?) scsi.device patches?

PFS3AIO shouldn't mount the partition if it crosses 4G (or ~8G if A600/A1200/A4000 IDE) border without required 64-bit support.

Okay I made a 10Gb dynamic hard file in winuae and did exactly the same thing to partition and format the drives. I then copied over my ADF collection totaling 4.7Gb to the 2nd partition of it first time with no problems at all. I did this using PFS3-AIO, loaded Classic WB up to partition it and copy the files to the hard file. I did not use idefix or change the scsi.device to get this done.

So I know my method works (well I knew that anyway since i've done other drives in the past) but it just will not work on physical drives without giving me block id errors. I get this problem with a 3.5" 8Gb WD Protege (taken out of an old xbox), a 2.5" 40Gb WD Scorpio and a 30Gb Fujitsu drive. Correct me if i'm wrong but WinUAE should be handling these drives fine without any scsi patching the same way it did with the hard file?

What other information do you need me to supply in order to give you a better idea?

[fitzsteve]

When you add the HDD if you select UAE as the device driver, then yes you are right there is no need to patch the scsi device.

If you choose IDE then you need to patch the scsi device just like on real Amiga. If you use Classic Workbench all you need to do is drop in the scsi.device to the 'Devs' drawer and Classic Workbench takes care of the rest (look in the LargeHD drawer inside 'MyFiles')

[-Acid-]

I know this thread is old now but I just thought I would let you all know I found the problem was the IDE to USB cable I was using to connect the drives to the PC. I replaced it with an IDE to Sata adapter and connected it to the main board and i've not had any errors since.

Thanks for everyone's input.

[crazyc]

Just to say I have spent a day or so tearing my hair out over this as I was using the latest PFS3 AIO etc.
I was using both the built in card reader and an external USB one and having same issue.
Wouldn't you know it though, I had plugged the USB card reader into the USB port that is part of the built in card reader - switching it to another nearby USB port seems to have solved the issue.