PS3 prepISO 1.33a strange behaviour in finding .ISO files on external drive formated with exFat

[Cheshire UK]

prepISO: 1.33a
webMAN MOD: 1.47.48k update 2
external drive: SSD 1TB
CFW: Evilnat 4.92


I am experiencing a strange behaviour with prepISO; it can't find some of the .ISO files in the extenal disk.
I've tried many things and the only one that seems to work is copying the file from within the PS3 using IRISMAN to the internal HDD, delete the file on the extrenal drive and copying the file back into the PS3ISO folder on the external HDD for those files the prepISO couldn't find - after doing this procedure, rescanning the system with preISO find the pesky files.

To clarify some checks I've performed (very short version of the tests):

1) It doesnt happen to all files, some of them - I couldn't find the pattern with which files this is happeing as per yet, but I believe files bigger than 8GB (not 100% sure yet)

2) I've checked if there where some strange special OS flags in the files or even some strange characters on their name
IRISMAN can see them, can rename them (I did to simple names and tried .iso and .ISO), can copy and delete them

3) the other suspition I have is the disk has only about 100GB free out of 1TB
I've got another 500GB SDD extrenal drive and the same files trnafered from the same computer don't have this issue (this has even less space left: only 90MB)
I even swap the case for the drivers (the "just in case is the controller doing something strange")

For to confirm, IRISMAN always list and can manipulate all files in the disk (rename,delete, copy) without issues



Perhaps this one is for aldostools (Aldo Vargas) - any ideas?

 

[aldostools]

Try defragmenting the ISO or the whole exFAT disk. If the ISO is too fragmented, the file will be ignored because the cache file has a maximum of 64KB and cannot store more than 8187 segments (sector start & number of continuous sectors that belong to the file). The cache file is created in /dev_hdd0/tmp/wmtmp and used by rawseciso and Cobra. rawseciso acts as an intermediary between the SCSI driver and Cobra to allow direct read access to the ISO on the storage device.

https://github.com/aldostools/webMAN-MOD/blob/master/_Projects_/prepISO/include/exfat.h#L291

 

[Cheshire UK]

I can certainly try that aldostools - any suggestion for a defrag that runs on Mac OSX ?

Out of interest (I haven't look into the code yet) is there a way to enable a bedug log that would show what is happening during the scan? like for instance "Skiping file ZZZZ - segments greater than 8187 for the file"

 

[aldostools]

I can certainly try that aldostools - any suggestion for a defrag that runs on Mac OSX ?

Out of interest (I haven't look into the code yet) is there a way to enable a bedug log that would show what is happening during the scan? like for instance "Skiping file ZZZZ - segments greater than 8187 for the file"

I'm not a Mac user. You could try this https://coriolis-systems.com/ but I don't know how effective it is.

There is not a debug version. However the source code is open, so you can make the changes that you want.

 

[Cheshire UK]

Thanks aldostools.

The Mac defrag I tried wasn't any good for exFat, I've installed "Defraggler" inside a Windows VM. (It wasn't ideal, I had to restart a few times, but got the files defragmented in the end).

The files were indeed very fragmented – on the screenshot you can see some that are still with a high fragmentation, particularly one with 14611 fragments (the second on the list) – however that particular file never was a problem for prepISO.

I was reflecting why those files were so fragmented (there were few with more then 1700 fragments) the answer was obvious: I copied them from the NAS appliance directly attaching the external SSD enclosure to it and I've trigger a copy of many files in parallel.
Said that, at the beginning (before recently when I just post in the forum) this issue never happened.

I believe that was it and it should not happen again – I will monitor the situation and if does even after a defrag, I will update the thread.

Once again thank you very much indeed for your input aldostools

Cheers

M.