PFS3AIO v3.1 Final

[RoC]

Apologies for posting this here. I am not sure whether the problem is, so I leave the admins to move wherever they feel appropriate.

I set-up a new HDF image, based on the excellent 16 GB V1.2 beta3 HstWB of Cyberstorm. The filesystem is the good PFS3 All-in-One v3 final.

http://hstwb.firstrealize.com/

The HDF works just fine within WinUAE, both by using the HDF and by launching the SD with WinUAE with Amiga 600/1200/4000 IDE.

PFS3AIO is the release v3 with identifier PFS/03 50465303 and transfer 0x1fe00.

Neverthelss, the SD card looks to be write-protected when launched on A1200, which has a SD/IDE adapter, ACA1230, Indivision MKIIa and kick 3.1 40.068

I managed to manually loadmodule the scsi.device 43.45 and loadwb from CLI and get a broken read-only workbench. Managed to launch HDToolBox and got the attached images.

I am clerly doing something wrong, however I don't know where the problem is..

Any advice would be appreciated to troubleshoot


To be tested:
- On Diskpart I can see a partition 1 of 15 Gigs on the SD card. I understand this shoulnd't be the case.

Tests already done:
- The same A1200 works with a 4 and 8 GB SD card with a ClassicWB image.
- Tested with both PFS/03 and PDS/03. The latter with 50445303
- The hardware SD "lock" is not locked. That was easy to check.
- From diskpart, checked the attribute write whether it was not on. Pushed the "attributes disk clear readonly" just to be on the safe-side.
- Cleaned the SD from diskpart several times, before proceeding. Formatted as Windows as a test and cleaned it back.
- Tested two different apps to flash the SD: WinImg 9 x64 and Win32DiskImager
- Tested two different SD adapters on two different PCs with Windows 10
- Tested three SD cards. Two new Kingstones (16 and 32 GB) and, importantly, one old SD card with a fully functioning ClassicWB on it. Once flushed it doesn't work, however I managed to flash back an old image and boot. It was SFS though.
- I have a SD cable extender. Removed it just as a test.
- I have a ACA 1230 accelerator. Removed it as a test.
- Created a -log file from WinUAE to facilitate the trobleshooting.

[Toni Wilen]

Error message means write attempt returned error code 3 from device driver (I am not sure what 3 means).

Any unexpected error causes PFS3 to enter read-only safe mode to prevent possible data corruption. It can be disabled with lock shell command.

If you run hdtoolbox, then change something, for example partition's number of buffers. Then save changes. Power cycle. Run hdtoolbox again. Did value change back to original?

It sounds like hardware problem. Do you have another power supply? (Solid state devices usually have protection that disables writes if they detect power problems) It is also possible adapter simply isn't compatible.

[RoC]

Originally Posted by Toni Wilen

Error message means write attempt returned error code 3 from device driver (I am not sure what 3 means).

Any unexpected error causes PFS3 to enter read-only safe mode to prevent possible data corruption. It can be disabled with lock shell command.

If you run hdtoolbox, then change something, for example partition's number of buffers. Then save changes. Power cycle. Run hdtoolbox again. Did value change back to original?

It sounds like hardware problem. Do you have another power supply? (Solid state devices usually have protection that disables writes if they detect power problems) It is also possible adapter simply isn't compatible.

Appreciated the clarifications Toni

Having tried another PSU, it didn't solve it.

Moreover, I can change the buffer from 30 to, say, 80 however when saving changes to disk within HDToolBox it returns the usual "Error 3 on write". I can change it by mounting on the WinUAE, however I don't believe it is going to help much.

I tried to use the lock command - good to learn about it - however lock dh0: off dind't do the magic.

What puzzle me it is that the ClassicWB works just fine on that A1200, though they are two SD cards flashed back in 2010 and smaller SDs (one 4 and one 8 Gigs)

Another point is the diskpart. I believe there shoulnd't be partitions visibles in the PFS3AIO's SD cards, however I can see one of 15 gigs. Is this an issue?

[Toni Wilen]

Originally Posted by roc

Moreover, I can change the buffer from 30 to, say, 80 however when saving changes to disk within HDToolBox it returns the usual "Error 3 on write". I can change it by mounting on the WinUAE, however I don't believe it is going to help much.

This confirms it is hardware problem. Either card or adapter. Or Amiga, some HDs simply won't work with Amiga IDE.

I tried to use the lock command - good to learn about it - however lock dh0: off dind't do the magic.

It is possible early error is considered too fatal.

What puzzle me it is that the ClassicWB works just fine on that A1200, though they are two SD cards flashed back in 2010 and smaller SDs (one 4 and one 8 Gigs)

Fake card? There are PC tests programs that fill the card completely and then validate data to detect fakes.

Another point is the diskpart. I believe there shoulnd't be partitions visibles in the PFS3AIO's SD cards, however I can see one of 15 gigs. Is this an issue?

Does other cards show same mysterious partition? If not, it sounds like card is doing something suspicious..

[RoC]

Originally Posted by Toni Wilen

This confirms it is hardware problem. Either card or adapter. Or Amiga, some HDs simply won't work with Amiga IDE. ..

Thanks for the help Toni.

I agree, It must be some kind of hw failure.

I have ordered a new adapter with a 2.5 cable, hoping that is the issue, but the A1200 could be problematic as well (hope not..but maybe it is time to recap).

EDIT:

Thanks to the Toni advice to look for HW issues, the problem has been solved. It was the combination of two issues:

• SD Adapter. For some reason the old one was not working well. I ordered a new one.
• SD Card. This was trivial, basically the new "super-duper" Kingston Micro-SD HC 32GB with its SD adapter was not working, That was easy to spoil, because the DH0 worked while DH1 didn't both on true A1200 and WinUAE. An old SD card without adapter worked it out. Tried to diskpart "clean" the 32 GB card, re-partition from HdToolBox leaving some unused space at the end of DH1, using pfsdoctor and finally testing different size of HDF images (8,16) without luck. It was showing DH1 as NDOS.

[Prosonic]

this last release doesn't flush the buffers fast enough im having a lot of issues of files / edits not saved

[Toni Wilen]

Originally Posted by Prosonic

this last release doesn't flush the buffers fast enough im having a lot of issues of files / edits not saved

Flush timing has not been changed.

[Hedeon]

Is there any way to recover files not in the deldir? Some nasty makefile error deleted not only #?.o but also #?.c files and the deldir does only contain a selection of the deleted files. :-(

[daxb]

With PFSsalv you can recover files and copy them to another disk. If not already done increase the deldir size for the next time.

[Hedeon]

Thank for the tip, but PFSSalv2 from Aminet is crashing hard on my MorphOS machine. There is no tool within PFS3AIO that can do this? Or be made part of PFSDoctor?

Originally Posted by daxb

With PFSsalv you can recover files and copy them to another disk. If not already done increase the deldir size for the next time.

[Toni Wilen]

http://www.winuae.net/files/b/pfs3aio_v3.1_test2.zip

WARNING: If you use OS 3.1.4 beta and selected new SCSI Direct option in HDToolBox when partitioning with v3.0 pfsaio: partition only mounts with v3.0 and v3.1 test #1. No other pfs3aio version can mount it. Partition should be reformatted with this version or without SCSI Direct option in HDToolBox. (If you don't, no support here)

PFS3AIO 3.1 Test #2

- OS 3.1.4 SCSIDIRECT option used SUPERFLOPPY drive total block count calculation/geometry mode even when SUPERFLOPPY was not set.
- Set default stack back to original 4000 (from old AROS too large value of 6000). Because system default stack is 4000, previously new stack was always allocated, wasting 6k of RAM.

(All previously discussed "small ram mode" changes temporarily removed, they may be back after 3.1 is out)

[tech3475]

Sorry for asking a basic question but how does this compare to SFS now?

When I try to look this up most of the information seems to be outdated but I'm in the middle of switching over to a mSD/CF adapter so now is the perfect time to change FS on my A1200.

So far it seems that PFS3 has the advantage of better OS support but apparently has a greater chance of data loss.

I do plan on having a >104GB partition, has anyone had any issues with this as it's listed as 'experimental'?

[Gilloo]

Originally Posted by tech3475

Sorry for asking a basic question but how does this compare to SFS now?

When I try to look this up most of the information seems to be outdated but I'm in the middle of switching over to a mSD/CF adapter so now is the perfect time to change FS on my A1200.

So far it seems that PFS3 has the advantage of better OS support but apparently has a greater chance of data loss.

I do plan on having a >104GB partition, has anyone had any issues with this as it's listed as 'experimental'?

PFS3aio intends to work with KS1.3 (even KS1.2) and 68000 and up.
That's why it's very useful to me (I would never thank enough Toni for his work)
For data loss PFSDoctor on <104GB partitions (from the official PFS3 on aminet http://aminet.net/package/disk/misc/PFS3_53) but is working only with KS3.1.
The data loss I encounted was after decompressing a big lha directly on the PFS3 partition. On copy, it never occured. this point has been fixed and never re-occur with PFS3aio...

[talybont]

Originally Posted by Prosonic

this last release doesn't flush the buffers fast enough im having a lot of issues of files / edits not saved

Is that because you reboot too soon after editing files? Maybe this will help: http://aminet.net/package/disk/salv/DiskSafe


I don't know if it works with PFS though.

[Hedeon]

I thought (original?) PFS3 used a keyboard reset handler to prevent this?

[tech3475]

Originally Posted by Gilloo

PFS3aio intends to work with KS1.3 (even KS1.2) and 68000 and up.
That's why it's very useful to me (I would never thank enough Toni for his work)
For data loss PFSDoctor on <104GB partitions (from the official PFS3 on aminet http://aminet.net/package/disk/misc/PFS3_53) but is working only with KS3.1.
The data loss I encounted was after decompressing a big lha directly on the PFS3 partition. On copy, it never occured. this point has been fixed and never re-occur with PFS3aio...

I might give it a shot then.

One downside is not having data recovery on >104GB partitions.

[kolla]

Originally Posted by tech3475

One downside is not having data recovery on >104GB partitions.

Always keep your data elsewhere too, sync with NAS or something.

[tech3475]

Originally Posted by kolla

Always keep your data elsewhere too, sync with NAS or something.

One of the reasons why I'm switching over to CF and ordered a rear access adapter.

But still it would be nice to have something just in case and/or reduce downtime.

[Toni Wilen]

Originally Posted by tech3475

One downside is not having data recovery on >104GB partitions.

First post has link to pfsdoctor version that supports >104G partitions.

[tech3475]

Originally Posted by Toni Wilen

First post has link to pfsdoctor version that supports >104G partitions.

Thanks, missed that.